1 00:00:09,210 --> 00:00:15,210 Jamie Duncan: Yes. Hey, and thanks for joining us on episode 2 00:00:15,210 --> 00:00:18,480 11 of the K files. My name is Jamie Duncan, I work at Google 3 00:00:18,480 --> 00:00:21,300 Cloud with me all, as always is tark Islam. 4 00:00:21,599 --> 00:00:24,179 Tariq Islam: Hey everyone, tarc Islam, also with Google Cloud, 5 00:00:25,260 --> 00:00:26,370 Jamie Duncan: and john Osborne, 6 00:00:27,239 --> 00:00:29,069 John Osborne: head runs is john Osborne with Red Hat. 7 00:00:29,969 --> 00:00:32,789 Jamie Duncan: JOHN is currently enjoying a Stella Artois from 8 00:00:32,789 --> 00:00:35,039 his home studio, and I'm a little jealous of him, 9 00:00:35,370 --> 00:00:36,120 John Osborne: well deserved. 10 00:00:38,010 --> 00:00:39,810 Jamie Duncan: It's Tuesday, and john said it's already been a 11 00:00:39,810 --> 00:00:46,410 run. So this is not going to be a coop con recap episode, 12 00:00:46,440 --> 00:00:49,590 because I think all three of us combined have had a chance to 13 00:00:49,590 --> 00:00:53,910 watch 17 minutes of coop con content. And since it ended last 14 00:00:53,910 --> 00:00:54,270 week, 15 00:00:56,070 --> 00:00:58,950 John Osborne: I tried. I don't know what you guys, I just, it's 16 00:00:59,550 --> 00:01:02,460 the whole idea of blocking out my calendar when I'm home. And 17 00:01:02,460 --> 00:01:05,340 then someone's like, I have this really high priority thing. And 18 00:01:05,340 --> 00:01:07,590 it's like, well, I'm sorry, I'm watching YouTube, like it 19 00:01:07,650 --> 00:01:11,670 doesn't, just doesn't, doesn't really work like when you're at 20 00:01:11,670 --> 00:01:13,710 it, as well as when you're actually physically at 21 00:01:13,710 --> 00:01:14,520 somewhere. So 22 00:01:14,670 --> 00:01:16,860 Jamie Duncan: yeah, yeah, I've run in I have this, I don't have 23 00:01:16,860 --> 00:01:20,760 the attention span for a virtual conference. I really don't. And 24 00:01:20,760 --> 00:01:24,150 Chris short, said it pretty well. He had a thread on 25 00:01:24,150 --> 00:01:27,810 Twitter, where he's kind of said, you know, this is my last 26 00:01:27,990 --> 00:01:31,560 virtual event. And he laid out some pretty good logic and some 27 00:01:31,560 --> 00:01:34,110 stuff I didn't wholly agree with. But I certainly understand 28 00:01:34,110 --> 00:01:36,600 where it's coming from. And I agree with the sentiment in 29 00:01:36,600 --> 00:01:41,130 totality. So So no, coop con recap. We're, I'm sure we will 30 00:01:41,130 --> 00:01:43,590 find as the content gets onto YouTube and gets out there. 31 00:01:43,590 --> 00:01:46,020 Well, I plan on watching a ton of it, there were a ton of good 32 00:01:46,020 --> 00:01:48,240 talks, but so far have watched almost none. 33 00:01:48,569 --> 00:01:51,029 John Osborne: I do feel like one thing I will say about coop con, 34 00:01:51,029 --> 00:01:54,509 I do feel like in the future, they'll probably have both modes 35 00:01:54,509 --> 00:01:57,869 where you can go there. But also it is in a good experience for 36 00:01:57,869 --> 00:02:02,159 people that want to be home. And it was cool. That was if you 37 00:02:02,159 --> 00:02:05,099 wanted to stream it it was you know, I think it was $100 or 38 00:02:05,099 --> 00:02:08,519 something which you know, right. And normally it's like $2,000 or 39 00:02:08,519 --> 00:02:11,909 something like that. And plus travel plus, travel. Yeah, all 40 00:02:11,909 --> 00:02:14,069 that stuff. So at least I made it Mark accessible to more 41 00:02:14,069 --> 00:02:14,669 people. But 42 00:02:14,760 --> 00:02:17,100 Jamie Duncan: yeah, making it democratizing in that way is 43 00:02:17,100 --> 00:02:19,410 really cool. But just being a virtual event was a little too 44 00:02:19,410 --> 00:02:19,740 rough. 45 00:02:20,190 --> 00:02:22,020 Tariq Islam: I think it comes down to delivery as well. I 46 00:02:22,020 --> 00:02:24,990 mean, for me, just you mentioning that travel and the 47 00:02:24,990 --> 00:02:28,170 time off and the hotel stay and then all that just exhausted me 48 00:02:29,640 --> 00:02:30,030 just 49 00:02:30,870 --> 00:02:33,330 Jamie Duncan: talking about it. But we are taking on something 50 00:02:33,330 --> 00:02:38,400 pretty ambitious. not going to get into the huge chunk of it, 51 00:02:38,400 --> 00:02:42,660 the short version is for the next several episodes, we are 52 00:02:42,660 --> 00:02:48,630 going to be tackling the majority of what ci CD means in 53 00:02:48,630 --> 00:02:52,200 the Kubernetes space today. And we're still settling some of the 54 00:02:52,200 --> 00:02:55,650 details. But it's going to be five or six episodes. And we 55 00:02:55,650 --> 00:02:58,380 have a question for people. And we'll probably throw a Twitter 56 00:02:58,380 --> 00:03:02,250 poll out there. JOHN will have details on that later. But this 57 00:03:02,250 --> 00:03:07,200 first episode, we're gonna get down I think first and foremost, 58 00:03:07,200 --> 00:03:12,180 we have to define what ci CD means. I always we always, we 59 00:03:12,180 --> 00:03:15,270 tend to start with definitions here. And then we're going to 60 00:03:15,270 --> 00:03:20,400 get into the backbones, the platforms on which you can do ci 61 00:03:20,400 --> 00:03:26,250 CD. And really kind of get down into the nuts and bolts a little 62 00:03:26,250 --> 00:03:29,100 bit and talk about some of the pros and the cons and what would 63 00:03:29,100 --> 00:03:35,730 we using today. So who wants to take a stab at defining ci CD? 64 00:03:37,410 --> 00:03:41,940 Tariq Islam: Let me start by saying that in the minds of many 65 00:03:41,940 --> 00:03:49,020 folks, you know, see ICD is sometimes conflated with DevOps. 66 00:03:49,380 --> 00:03:54,420 And I think it's important for us to start defining c ICD by 67 00:03:54,450 --> 00:03:59,730 first defining what it's not. And see, ICD is definitely not 68 00:03:59,790 --> 00:04:03,990 DevOps, but it certainly falls under the auspices of DevOps. I 69 00:04:03,990 --> 00:04:06,930 think I think we'll start teasing out DevOps through the 70 00:04:06,930 --> 00:04:11,910 course of these episodes. Because they are inherently 71 00:04:11,910 --> 00:04:16,350 interrelated. And again, I think ci CD represents a subset 72 00:04:16,350 --> 00:04:21,390 manifestation of a DevOps dev sec Ops, Dev, appsec, ham 73 00:04:21,390 --> 00:04:27,840 sandwich. We put it culture, but I think, I think I think it's 74 00:04:27,840 --> 00:04:30,870 important again, for the listeners to just settle, you 75 00:04:30,870 --> 00:04:34,320 know, the backdrop here is that they are, they are very separate 76 00:04:34,320 --> 00:04:39,300 topics to discuss. And with C ICD, and how we're going to 77 00:04:39,300 --> 00:04:44,640 approach it, it is about the tools. But you know, and john 78 00:04:44,640 --> 00:04:48,990 made this great point earlier. It's also about the process 79 00:04:49,020 --> 00:04:53,010 around those tools and and the impact that all of that has on 80 00:04:53,010 --> 00:04:56,070 the people. So you can start seeing where that bleeds into 81 00:04:56,070 --> 00:05:00,990 the DevOps part of it. But it's not it's not everything. DevOps. 82 00:05:01,110 --> 00:05:04,020 I know that sounds weird, but I think it'll become clearer as we 83 00:05:04,050 --> 00:05:05,430 as we go through these episodes. 84 00:05:05,850 --> 00:05:07,560 John Osborne: Yeah, I mean, if I always tell people like if 85 00:05:07,560 --> 00:05:10,020 DevOps was just purely a technical problem, it already 86 00:05:10,020 --> 00:05:14,070 would have been solved, right? The hard part is the people. But 87 00:05:14,100 --> 00:05:18,990 I think at the same time, if you look at what you can do in an 88 00:05:18,990 --> 00:05:22,920 organization for most people, it is they can automate something. 89 00:05:23,280 --> 00:05:26,310 And that's why we I think, why we end up talking about ci CD so 90 00:05:26,310 --> 00:05:30,300 much, because it is a, it's more tangible in the sense that more 91 00:05:30,300 --> 00:05:32,460 people can make a difference through automation, and some of 92 00:05:32,460 --> 00:05:36,030 these some of these practices, whereas, you know, defining 93 00:05:36,420 --> 00:05:41,310 outcomes, and you know, reorg, doing reorg is when necessary, 94 00:05:41,310 --> 00:05:44,100 and some of the other cross functional team stuff ends up 95 00:05:44,100 --> 00:05:45,990 being, you know, there's only a small number of people in 96 00:05:45,990 --> 00:05:48,480 organizations that can change some of those things that you 97 00:05:49,440 --> 00:05:54,030 that ended up need to be changed at some point. So 98 00:05:55,650 --> 00:05:57,210 Jamie Duncan: yes, it sounds like we're gonna be talking a 99 00:05:57,210 --> 00:06:00,600 lot about technical debt from a couple of different angles. 100 00:06:01,050 --> 00:06:02,280 Tariq Islam: I think we have to Yep. 101 00:06:02,610 --> 00:06:04,890 Jamie Duncan: Yeah. And it's kind of interesting. In I had 102 00:06:04,890 --> 00:06:07,230 this discussion, I forget who I was talking to whether it was on 103 00:06:07,230 --> 00:06:11,340 Twitter or in like an internal chat room or what. But um, 104 00:06:12,570 --> 00:06:15,750 technical debt takes on this negative connotation 105 00:06:15,750 --> 00:06:20,220 immediately. But when you're are doing something like john said, 106 00:06:20,220 --> 00:06:22,380 when you are making these positive changes, because if 107 00:06:22,380 --> 00:06:24,750 you're a completely manual shop that runs off with three ring 108 00:06:24,750 --> 00:06:29,100 binders, and three gay and conference calls, going to 109 00:06:29,130 --> 00:06:32,430 getting into, you know, automating your pipelines for 110 00:06:32,430 --> 00:06:38,070 continuous deployment is taking on new technical debt. And there 111 00:06:38,070 --> 00:06:40,920 are good kinds of technical debt to take on, you know, it's like, 112 00:06:41,400 --> 00:06:44,580 there is, you know, there's payday loan technical debt of 113 00:06:44,580 --> 00:06:47,670 people making poor choices, and everyone having to endure them, 114 00:06:47,790 --> 00:06:50,940 you know, for a long time, organizationally. But there's 115 00:06:50,940 --> 00:06:55,260 also mortgage technical debt, where the debt is, you know, of 116 00:06:55,260 --> 00:06:58,410 a low enough interest rate to push the to push it a little too 117 00:06:58,410 --> 00:07:02,250 far. But it It provides a lot of benefit for what you're taking 118 00:07:02,250 --> 00:07:07,830 the dead on for, and ci CD, a CD platform. Most of the time, 119 00:07:08,160 --> 00:07:12,390 unless it's just a really poor implementation is good technical 120 00:07:12,390 --> 00:07:13,200 debt to take on. 121 00:07:13,980 --> 00:07:16,230 John Osborne: Yeah, I think too, like, you know, talking about 122 00:07:16,230 --> 00:07:20,010 technical debt. You think about agile and some other things that 123 00:07:20,010 --> 00:07:23,430 also, you know, there's a Venn diagram somewhere out there, but 124 00:07:23,430 --> 00:07:27,390 that get entwined with things like ci CD sometimes in Sprint's 125 00:07:27,390 --> 00:07:31,470 and these things, and there is almost this feeling out there. I 126 00:07:31,470 --> 00:07:35,160 think that's honestly probably justified that a lot of times, 127 00:07:35,160 --> 00:07:38,430 we get so focused on features that we're not adequately 128 00:07:38,910 --> 00:07:41,520 prioritizing. I'm totally gonna go off the rails in this 129 00:07:41,520 --> 00:07:46,230 episode. But we're not prior, we're not prioritizing our 130 00:07:46,230 --> 00:07:50,280 technical debt and our testing and good architecture diagrams 131 00:07:50,280 --> 00:07:53,010 and documentation and those things because you start doing 132 00:07:53,010 --> 00:07:56,490 two week sprints and things like that it gets it does get, if you 133 00:07:56,490 --> 00:07:58,950 do that poorly, it actually can be worse than waterfall if you 134 00:07:58,950 --> 00:08:03,720 don't actually start. You know, prioritizing things the way they 135 00:08:03,720 --> 00:08:06,060 should. And I will say to for as a side tangent, we'll put it in 136 00:08:06,060 --> 00:08:09,180 the show notes, there's a new really good book out called 137 00:08:09,180 --> 00:08:12,660 project to product on a flow framework. It's this kind of 138 00:08:12,660 --> 00:08:17,640 prescriptive plan for value stream management. And the the 139 00:08:18,120 --> 00:08:21,030 author also spoke at the DevOps enterprise summit, He's really 140 00:08:21,030 --> 00:08:25,290 good. I've never tried this, but just reading it and in watching 141 00:08:25,290 --> 00:08:29,730 it, it makes sense and resonated with me a little bit in terms 142 00:08:29,730 --> 00:08:33,330 of, you know, how much should you how much of your Sprint's 143 00:08:33,330 --> 00:08:38,070 should you focus on things like reducing your technical debt and 144 00:08:38,430 --> 00:08:41,760 new features? And what makes and when should you adjust that and 145 00:08:41,760 --> 00:08:44,520 why and it all sort of made sense to me? Yeah, I 146 00:08:44,520 --> 00:08:46,260 Jamie Duncan: mean, the traditional, you know, the old 147 00:08:46,260 --> 00:08:50,310 trope is that features drive revenue. But even with 148 00:08:50,310 --> 00:08:54,120 Kubernetes, we've gotten to the point where like cube one died, 149 00:08:54,120 --> 00:08:57,330 18 was very much a bug fix release, there were very few new 150 00:08:57,330 --> 00:09:00,390 features, very few things moved from alpha to beta, or beta to 151 00:09:00,390 --> 00:09:02,940 GA. And like you're talking about john, like, when you're in 152 00:09:02,940 --> 00:09:07,890 these two week sprints, you can fixate on the wrong things, and 153 00:09:07,890 --> 00:09:10,260 really drive your product into an unhealthy state. 154 00:09:10,409 --> 00:09:13,139 Tariq Islam: So I guess then, the goal of properly done CIC D 155 00:09:13,139 --> 00:09:18,119 is to avoid, you know, forgetting the forest. Right, 156 00:09:18,149 --> 00:09:21,899 and only focusing on the trees. Yeah. And that's, and I like 157 00:09:21,899 --> 00:09:25,649 your call out about technical debt, Jamie. But I do I do have 158 00:09:25,649 --> 00:09:29,879 a recent story. Actually, I was working with an organization 159 00:09:29,939 --> 00:09:36,359 that they were still packaging up now actually, to this day, 160 00:09:36,479 --> 00:09:41,189 they are still packaging up their, their binaries in a zip 161 00:09:41,189 --> 00:09:43,859 folder and emailing them to their ops team for deployment. 162 00:09:44,070 --> 00:09:45,480 John Osborne: I've seen that all the time. So 163 00:09:45,870 --> 00:09:47,700 Tariq Islam: yeah, it was Yeah, 164 00:09:47,820 --> 00:09:49,500 John Osborne: I just moving it around, you know. 165 00:09:50,669 --> 00:09:52,499 Jamie Duncan: Not even SFTP Yeah, not even. 166 00:09:54,270 --> 00:09:56,520 Tariq Islam: Yeah, I see that all the time. So but but that's 167 00:09:56,520 --> 00:10:02,670 where folks are right. And I think i think i think CIC D it 168 00:10:02,730 --> 00:10:08,340 to John's point, it really does require a cultural shift just 169 00:10:08,340 --> 00:10:11,880 like Kubernetes adoption does that we've talked about before. 170 00:10:11,940 --> 00:10:16,320 I mean, this is we're talking about not just automation to 171 00:10:16,320 --> 00:10:19,920 solve problems, but this is this is, you know, organizational 172 00:10:19,920 --> 00:10:22,620 infrastructure that you're going to build on top of. 173 00:10:23,130 --> 00:10:24,750 John Osborne: I'm really getting off the rails In this episode, 174 00:10:24,750 --> 00:10:26,640 by the way, I feel will 175 00:10:26,640 --> 00:10:27,930 Jamie Duncan: keep you on track, we'll keep you on 176 00:10:29,429 --> 00:10:31,379 John Osborne: Kubernetes like actually makes a lot of sense 177 00:10:31,379 --> 00:10:33,749 here. And there's other technologies too, but you know, 178 00:10:33,779 --> 00:10:37,679 it's, we talk about, I think, Craig Mcluckie talked about this 179 00:10:37,679 --> 00:10:40,319 like last year, but you know, the move from ticket driven to 180 00:10:40,319 --> 00:10:43,109 API driven well, instead of putting in a ticket for saying 181 00:10:43,109 --> 00:10:46,259 you need, what storage and how many replicas you want, you 182 00:10:46,259 --> 00:10:49,619 know, that can be part of the API that developers hand over to 183 00:10:49,619 --> 00:10:52,019 operations teams where they can say, here's my deployment, I 184 00:10:52,019 --> 00:10:56,729 need, I need, you know, this persistent volume and need to 185 00:10:56,729 --> 00:10:58,319 run this number of replicas within this. 186 00:10:58,740 --> 00:11:00,540 Tariq Islam: So that's an interesting point. Because 187 00:11:00,570 --> 00:11:04,530 Because I don't think there's a singular formula for ci CD. I 188 00:11:04,530 --> 00:11:07,680 mean, there's like, okay, build your app, run your test, create 189 00:11:07,680 --> 00:11:11,220 your artifact, you know, deploy to staging, you know, yada yada, 190 00:11:11,220 --> 00:11:13,770 yada deploy to production, Canary, whatever, and then 191 00:11:13,770 --> 00:11:18,000 you're good to go. But that can manifest itself in many 192 00:11:18,000 --> 00:11:23,490 different ways, where different organizations define the lines 193 00:11:23,490 --> 00:11:27,870 between their teams differently. And I've seen ci CD work where 194 00:11:27,870 --> 00:11:31,290 developers are reaching over to the right, where they are 195 00:11:31,350 --> 00:11:39,930 dealing with, you know, API driven requests, or that that 196 00:11:39,930 --> 00:11:42,930 engagement pattern is definitely something that I've seen work, 197 00:11:42,930 --> 00:11:46,500 but it requires a lot of, I think, a high degree of maturity 198 00:11:46,500 --> 00:11:48,600 that most probably don't have. 199 00:11:48,840 --> 00:11:50,430 Jamie Duncan: And there's a there's a pretty important 200 00:11:50,430 --> 00:11:54,810 distinction as well, that I want to draw in to try to segue this 201 00:11:54,810 --> 00:11:58,350 into talking about sort of how you get started, like talking 202 00:11:58,380 --> 00:12:01,620 like using target customer, you know, whose zip file, you know, 203 00:12:02,220 --> 00:12:05,640 emailing zip files, to ops people to deploy to deploy new 204 00:12:05,640 --> 00:12:09,180 versions of things. Were john, where you pointed out, you see 205 00:12:09,180 --> 00:12:12,450 the principles and Kubernetes very much our peanut butter and 206 00:12:12,450 --> 00:12:14,670 jelly kind of scenario, chocolate and peanut butter. 207 00:12:15,060 --> 00:12:17,640 Apparently, peanut butter goes with everything, really. But 208 00:12:18,419 --> 00:12:20,909 Tariq Islam: it really including me, oh, by the way, believe it 209 00:12:20,909 --> 00:12:21,269 or not, 210 00:12:21,360 --> 00:12:24,450 Jamie Duncan: yeah, I've seen it, and not in that camp. But 211 00:12:24,450 --> 00:12:27,060 let's just take a second to recognize the universal love of 212 00:12:27,060 --> 00:12:31,200 peanut butter. But there, the there is a big difference, 213 00:12:31,200 --> 00:12:35,730 because like you can't be a little bit Kubernetes, you can 214 00:12:35,730 --> 00:12:39,000 be a little bit pregnant. Once you stand up a Kubernetes 215 00:12:39,000 --> 00:12:44,040 cluster. To deploy an application effectively, you've 216 00:12:44,040 --> 00:12:47,490 got to do all of it, you've got to containerize it, you've got 217 00:12:47,490 --> 00:12:49,860 to figure out config maps, you've got to get your ingress 218 00:12:49,860 --> 00:12:53,700 and your networking and your storage squared away, you can't 219 00:12:53,700 --> 00:12:56,610 do a little bit of Kubernetes, you don't have to put all of 220 00:12:56,610 --> 00:12:59,880 your applications on Kubernetes. But to do our application on 221 00:12:59,880 --> 00:13:04,230 cube, you got to do it all the way. CDs a little different, 222 00:13:04,410 --> 00:13:08,280 because that team that is you know, receiving zip files in 223 00:13:08,280 --> 00:13:13,470 their inbox to do to run a maintenance event. They can 224 00:13:13,980 --> 00:13:18,420 start down a road and and CD principals can be taken on one 225 00:13:18,420 --> 00:13:24,330 or five at a time. And all of them provide benefit. You don't 226 00:13:24,330 --> 00:13:27,810 have to go from zero to 100, like you do with an app in coop. 227 00:13:28,290 --> 00:13:32,940 You can build it up over time and drive that adoption drive 228 00:13:32,940 --> 00:13:34,860 those cultural changes with the tooling. 229 00:13:35,309 --> 00:13:36,749 Tariq Islam: And that's a good point. Because I think that 230 00:13:36,779 --> 00:13:41,819 that's why I've we have seen organizations adopt c ICD far 231 00:13:41,819 --> 00:13:45,239 before, or long before they adopt Kubernetes in that 232 00:13:45,719 --> 00:13:49,079 holistic way. Right. And by the way, I'm totally stealing that 233 00:13:49,109 --> 00:13:50,819 whole you can't be a little bit pregnant. 234 00:13:51,300 --> 00:13:52,380 Jamie Duncan: Oh, you never heard that before. 235 00:13:52,409 --> 00:13:53,729 Tariq Islam: I've never heard that one before. 236 00:13:53,760 --> 00:13:56,850 John Osborne: Oh, are you serious? I've never get around 237 00:13:56,850 --> 00:13:57,360 it for a long 238 00:13:58,740 --> 00:14:00,570 Tariq Islam: time. Michelle, I'm sheltered. What can you say? 239 00:14:02,399 --> 00:14:03,179 Jamie Duncan: Poor little Tarik, 240 00:14:03,179 --> 00:14:05,069 John Osborne: you're a nice boy. I'm telling you, man. I'm 241 00:14:05,070 --> 00:14:06,090 Tariq Islam: a nice boy, you know. 242 00:14:07,440 --> 00:14:09,960 Jamie Duncan: So let's, let's talk a little bit about some of 243 00:14:09,960 --> 00:14:13,020 the options. You know, this first episode we want to talk 244 00:14:13,020 --> 00:14:15,900 about the definitions. Before we get to that I want to get into a 245 00:14:15,900 --> 00:14:19,680 little bit of trivia with you too. And I it's one of those 246 00:14:19,680 --> 00:14:22,140 that where I'm wondering if I if I was sheltered back in the day, 247 00:14:22,470 --> 00:14:24,930 when the first place I was at that was doing anything 248 00:14:24,930 --> 00:14:28,410 resembling a CI CD workflow, and they were doing it really 249 00:14:28,410 --> 00:14:33,090 poorly. But I didn't know any better at the time. It was 250 00:14:33,090 --> 00:14:37,770 continuous integration. Continuous Delivery, was what ci 251 00:14:37,770 --> 00:14:41,010 CD stood for. Apparently now it's continuous integration, 252 00:14:41,010 --> 00:14:42,210 continuous deployment. 253 00:14:42,570 --> 00:14:43,800 Tariq Islam: I think that's still debatable. 254 00:14:44,069 --> 00:14:46,739 Jamie Duncan: Is it so is delivery Am I am I the only is 255 00:14:46,739 --> 00:14:50,699 that like this. Sinbad being the genie in the movie thing? Yeah. 256 00:14:51,840 --> 00:14:55,290 John Osborne: Just you know, deployment to multiple I guess 257 00:14:55,290 --> 00:14:58,020 if you're saying delivering it. I think it I think they changed 258 00:14:58,020 --> 00:14:59,880 it when they started. The delivery started happening to 259 00:14:59,880 --> 00:15:03,990 me. multiple targets. And okay, then they just started moving 260 00:15:03,990 --> 00:15:06,510 into deployment. That was kind of my, my take on it. 261 00:15:06,720 --> 00:15:09,300 Tariq Islam: I'm still a fan of delivery. Actually, I feel like 262 00:15:09,300 --> 00:15:10,560 it's a little bit more generic. 263 00:15:10,649 --> 00:15:12,749 Jamie Duncan: And I wonder if I wonder if there was like a book 264 00:15:12,749 --> 00:15:14,819 written. I wonder if there's, I wonder if there are camps here 265 00:15:14,819 --> 00:15:17,309 and we don't know about it. There are some blogs that talk 266 00:15:17,309 --> 00:15:19,979 about the difference between delivery and deployment. I'll be 267 00:15:19,979 --> 00:15:20,669 honest with you, 268 00:15:21,479 --> 00:15:22,349 Tariq Islam: I don't think it matters. 269 00:15:23,880 --> 00:15:25,740 John Osborne: I remember the first time that that's our it's 270 00:15:25,740 --> 00:15:26,070 like, 271 00:15:26,159 --> 00:15:27,089 Jamie Duncan: Reddit hated it. 272 00:15:28,890 --> 00:15:31,740 John Osborne: The first organization I worked for that 273 00:15:31,740 --> 00:15:35,400 did this. I was the first few years I was in the industry. I 274 00:15:35,400 --> 00:15:40,740 just wrote code by myself in a in in isolation, and Tareq. And 275 00:15:40,740 --> 00:15:43,170 I were joking about a month ago about how ahead some of these 276 00:15:43,170 --> 00:15:45,330 people are they're like, speaking on stage at cube con 277 00:15:45,330 --> 00:15:49,080 and stuff for over them. We were but the first organization I 278 00:15:49,080 --> 00:15:51,390 went to that actually did anything like see ICD, honestly, 279 00:15:51,390 --> 00:15:56,160 I, I was very probably immature. And I wasn't used to doing all 280 00:15:56,160 --> 00:15:58,950 these things. And I remember I went down to this, this guy's 281 00:15:58,950 --> 00:16:01,020 desk, and he's like, yeah, you know, we're gonna, we're gonna 282 00:16:01,020 --> 00:16:03,180 check our code. And every night, we're gonna scan it with a run 283 00:16:03,180 --> 00:16:05,010 these tests, if things break, we're gonna email you right 284 00:16:05,010 --> 00:16:09,630 away. And I just remember thinking, chill out, man. I just 285 00:16:09,630 --> 00:16:13,770 want to write some code. I don't want emails every night with 286 00:16:13,770 --> 00:16:15,960 things breaking, because things always break. And you know, I 287 00:16:15,960 --> 00:16:19,170 didn't, I didn't actually see later I moved into ops. And so I 288 00:16:19,170 --> 00:16:23,100 kind of saw all of why that mattered. But when I was just in 289 00:16:23,100 --> 00:16:26,340 that writing code mindset, I didn't really didn't see, at 290 00:16:26,340 --> 00:16:29,250 that time, at least, while that stuff was so important. 291 00:16:29,549 --> 00:16:31,829 Jamie Duncan: I remember getting all of those alerts. And they 292 00:16:31,829 --> 00:16:34,289 were constant. And they were useless, because they were 293 00:16:34,289 --> 00:16:37,169 poorly configured. Because we had no clue what we were doing. 294 00:16:37,709 --> 00:16:40,379 And they just became white noise. And you ignored them. 295 00:16:41,129 --> 00:16:41,489 That's, 296 00:16:41,520 --> 00:16:43,320 John Osborne: that's what happens still happens. There is 297 00:16:43,320 --> 00:16:45,960 no technology that solves that problem. I do think you need to 298 00:16:45,990 --> 00:16:48,300 there's a, there's just the data you got to put the time in. 299 00:16:49,169 --> 00:16:51,689 Tariq Islam: It's a it's a it's a maturity thing, right. It's an 300 00:16:51,689 --> 00:16:54,779 organization learning about what it cares about, in terms of 301 00:16:54,779 --> 00:16:58,049 delivering features and functions. It gets better over 302 00:16:58,049 --> 00:17:01,109 time. But I think, and john, I think you mentioned this before 303 00:17:01,109 --> 00:17:06,599 we started up, you know, how you measure your ci CD pipelines, 304 00:17:06,599 --> 00:17:10,289 your automation infrastructure, that's, that's hugely important. 305 00:17:10,529 --> 00:17:12,839 And you're not going to get it right the first time, you're not 306 00:17:12,839 --> 00:17:14,909 going to get it right, the eighth time, you might get it 307 00:17:14,909 --> 00:17:19,169 right, maybe the 18th time. But I think it's, it's part of that, 308 00:17:19,529 --> 00:17:23,459 that ci CD journey. And you have to have the patience and the 309 00:17:23,459 --> 00:17:27,299 discipline to kind of stick with it, and get to a point where 310 00:17:27,929 --> 00:17:30,929 it's not just about the efficiency of the automation, 311 00:17:30,929 --> 00:17:35,699 but the the data that you're getting back. Is it useful? And 312 00:17:35,699 --> 00:17:39,119 is it? Is it making your processes better? Is it 313 00:17:39,119 --> 00:17:39,839 impactful? 314 00:17:39,990 --> 00:17:42,360 John Osborne: I mean, the general kind of pattern I see 315 00:17:43,230 --> 00:17:48,060 is, organizations that, you know, initially, they really 316 00:17:48,060 --> 00:17:50,340 just talk about inputs, you know, we've got to, we've got to 317 00:17:50,340 --> 00:17:54,000 do cloud, we've got to do agile, you know, we've got to, we've 318 00:17:54,000 --> 00:17:56,880 got we need the Kubernetes, given the Kubernetes. And then, 319 00:17:57,810 --> 00:18:00,150 you know, once they get a little more mature, they start looking 320 00:18:00,150 --> 00:18:03,420 at the outputs, like the, you know, the four key metrics from 321 00:18:03,840 --> 00:18:06,690 the state of DevOps reports, the number of incidents, those 322 00:18:06,690 --> 00:18:09,720 things, and then once the few organizations have had that have 323 00:18:09,720 --> 00:18:13,890 gotten pretty mature, we'll talk about outcomes. So you know, 324 00:18:13,890 --> 00:18:17,610 reducing costs and improving operations teams and those types 325 00:18:17,610 --> 00:18:21,000 of things that actually make more of a difference. But 326 00:18:21,000 --> 00:18:23,640 initially, it's like, you know, we've got a, we need the cloud. 327 00:18:23,670 --> 00:18:26,850 I mean, it'd be the cloud. And but measurements do become 328 00:18:26,850 --> 00:18:30,180 really important, I think, in general to, to understanding 329 00:18:30,180 --> 00:18:30,870 success. 330 00:18:31,320 --> 00:18:34,260 Tariq Islam: So this, this begs the question, then, you know, 331 00:18:34,290 --> 00:18:38,100 what, what is it good? What is it good ci CD infrastructure 332 00:18:38,700 --> 00:18:46,350 look like beyond just build test? stage, deploy, monitor. 333 00:18:46,680 --> 00:18:49,440 John Osborne: So before Jamie answers to like, let me just 334 00:18:49,440 --> 00:18:51,600 throw, let me just throw everything we did in the, in the 335 00:18:51,600 --> 00:18:57,450 pre, the pre call of the window, cuz I want to bring something 336 00:18:57,450 --> 00:18:59,400 up, see if this resonates with you guys. But there is a really 337 00:18:59,400 --> 00:19:04,020 good article on scaling to 100,000 users, where they this, 338 00:19:04,080 --> 00:19:06,030 I'll put it in the show notes. And it's nothing to do with ci 339 00:19:06,030 --> 00:19:08,010 CD. But I thought we could actually use a similar format 340 00:19:08,040 --> 00:19:11,550 for this episode is based off what Sam and Jamie say, kind of 341 00:19:11,580 --> 00:19:14,520 jog my memory, but they kind of talked about, you know, building 342 00:19:14,520 --> 00:19:17,430 something with one user on one machine and then going into 10 343 00:19:17,430 --> 00:19:20,160 users and you split up the database and then 100 users and 344 00:19:20,160 --> 00:19:23,700 then building up like, maybe we do want to work that way for a 345 00:19:23,700 --> 00:19:26,010 CI CD, or we're talking about you know, if it's Let's just eat 346 00:19:26,010 --> 00:19:28,440 what should you do and kind of build out to what is like a 347 00:19:28,440 --> 00:19:31,260 large enterprise, but what should they be doing? Because I 348 00:19:31,260 --> 00:19:33,090 do think that the answer differs. 349 00:19:33,630 --> 00:19:36,840 Jamie Duncan: Yeah, if it's just me, my Git commits are going to 350 00:19:36,840 --> 00:19:37,980 be purely curse words. 351 00:19:40,320 --> 00:19:42,120 John Osborne: You can't look at my I have nothing. I have no 352 00:19:42,120 --> 00:19:43,080 documentation on code. 353 00:19:43,470 --> 00:19:45,870 Jamie Duncan: Right. Documentation will not exist. 354 00:19:45,930 --> 00:19:50,700 Yeah. I think I like the idea of starting with one person. I 355 00:19:50,700 --> 00:19:53,040 don't think it's particularly practical. He's you. It's one of 356 00:19:53,040 --> 00:19:55,290 those, you know, it's a Dilbert joke. 357 00:19:55,889 --> 00:19:57,329 Tariq Islam: You know, we're like, have you started working 358 00:19:57,329 --> 00:20:00,749 yet? And no, I've spent seven days configuring My environment. 359 00:20:00,749 --> 00:20:04,679 But but but let's go back to that example of of the 360 00:20:04,679 --> 00:20:09,179 organization that still sending zip files in an email to the SIS 361 00:20:09,179 --> 00:20:12,869 ops team that probably started with a single person somewhere. 362 00:20:13,019 --> 00:20:15,809 Yeah, sure. All right. And they didn't try to scale around that. 363 00:20:16,319 --> 00:20:17,729 I think that's a good approach. 364 00:20:18,779 --> 00:20:20,519 John Osborne: I think what happens when you see those 365 00:20:20,519 --> 00:20:24,209 things a lot of times is the hire at the top of the 366 00:20:24,209 --> 00:20:27,269 organizations don't have any idea who's talking to each 367 00:20:27,269 --> 00:20:31,739 other. And so what happens is, is on the ground floor, you 368 00:20:31,739 --> 00:20:34,619 know, they needed to get this to the operations teams, and no one 369 00:20:34,619 --> 00:20:37,559 really thought about silos talking to each other. So how do 370 00:20:37,559 --> 00:20:40,379 you contact someone, especially, you know, all these things 371 00:20:40,409 --> 00:20:43,079 process around for 20 years? So how did you how did you get 372 00:20:43,079 --> 00:20:45,599 something similar? 20 years ago, you emailed it right? And so 373 00:20:46,289 --> 00:20:50,399 nobody really cared to focus on doing an improvement of that. So 374 00:20:50,399 --> 00:20:53,489 here we are, was, that's still our problem. Years later, yeah, 375 00:20:53,489 --> 00:20:54,419 20 years later. Still, 376 00:20:54,449 --> 00:20:56,609 Tariq Islam: I think I think one of the best ways to distill this 377 00:20:56,609 --> 00:21:01,019 maybe is is looking at Conway's Law, perhaps because I think 378 00:21:01,019 --> 00:21:04,769 cicd flies in the face of that a little bit. where, you know, 379 00:21:04,769 --> 00:21:07,739 Conway's Law states that an organization is doomed or 380 00:21:07,739 --> 00:21:12,059 destined, pick your word of choice there to to create 381 00:21:12,359 --> 00:21:15,569 systems and software that mimic the communication patterns of 382 00:21:15,569 --> 00:21:22,109 their, of their organization. See, ICD is at least properly 383 00:21:22,109 --> 00:21:26,069 done is actually meant to break that or improve upon that. And I 384 00:21:26,069 --> 00:21:31,019 think that's, that's what makes it hard to get right. But to 385 00:21:31,079 --> 00:21:34,919 Jamie's point, there are areas of ci and CD, particularly on 386 00:21:34,919 --> 00:21:39,959 the CI CD part where actually on both sides, where you can, you 387 00:21:39,959 --> 00:21:44,159 can cherry pick a little bit on what you want to improve upon 388 00:21:44,189 --> 00:21:48,929 before taking a bigger bite. And so I think I think, you know, if 389 00:21:48,929 --> 00:21:52,289 an organization wants to do things better, it's a great 390 00:21:52,289 --> 00:21:55,739 place to start. It's a it's a good path of least resistance. 391 00:21:56,219 --> 00:21:59,069 And and once you really get really good at C, ICD and 392 00:21:59,069 --> 00:22:03,059 automation, I think, I think appetites increase for more 393 00:22:03,059 --> 00:22:07,529 advanced tooling and platforms, you know, Kubernetes, and the 394 00:22:07,529 --> 00:22:08,009 like. 395 00:22:08,220 --> 00:22:09,900 Jamie Duncan: So let's talk about getting started a little 396 00:22:09,900 --> 00:22:13,080 bit, because you have to, you know, see the end, we were 397 00:22:13,080 --> 00:22:15,900 talking a little bit about this, before we hit the record button, 398 00:22:15,930 --> 00:22:20,220 CD is the tooling. And, you know, that's really where the 399 00:22:20,220 --> 00:22:22,680 the sort of the line of demarcation begins, but between 400 00:22:22,680 --> 00:22:25,680 defining ci CD and defining DevOps, and they are very 401 00:22:25,710 --> 00:22:29,580 knitted together, but see ICD very much represent a toolset 402 00:22:29,580 --> 00:22:32,580 that you've adopted to get work done. Whereas DevOps is much 403 00:22:32,580 --> 00:22:37,590 more philosophy. If you're adopting, taking that idea of 404 00:22:37,590 --> 00:22:41,190 torrox example of the company that sends a zip file for over 405 00:22:41,190 --> 00:22:47,550 email for deployment jobs. If the, if the visionary in that 406 00:22:47,550 --> 00:22:51,180 group, we're starting to build out some sort of CD pipeline, 407 00:22:52,230 --> 00:22:55,620 they would have to start with an engine, they would have to start 408 00:22:55,620 --> 00:22:59,010 with a whatever you want to call it a backbone, a platform, an 409 00:22:59,010 --> 00:23:04,200 engine to begin to integrate the CD steps. And it's that's the 410 00:23:04,200 --> 00:23:07,320 first step they have to get on top of they have to take that 411 00:23:07,320 --> 00:23:08,700 first step up. 412 00:23:08,790 --> 00:23:11,910 Tariq Islam: Yeah, and and I'll go ahead and call this first one 413 00:23:11,910 --> 00:23:17,610 out. And but before I call it out, I am going to call back to 414 00:23:18,180 --> 00:23:21,300 a previous episode when we were talking about remember when we 415 00:23:21,300 --> 00:23:25,260 talked about how Helm benefited from being kind of first to 416 00:23:25,260 --> 00:23:29,850 land, and then being shoehorned for a number of use cases that 417 00:23:29,880 --> 00:23:34,830 aren't so great anymore? Right. I think that's that's where a 418 00:23:34,830 --> 00:23:38,280 lot of people landed with Jenkins. Jenkins has been around 419 00:23:38,520 --> 00:23:45,210 forever. And it's a build server. Right? It it since 420 00:23:45,210 --> 00:23:48,780 inception. And it's a very good one. But it's, it was almost 421 00:23:48,780 --> 00:23:52,890 purely ci and because of its pluggable architecture, it 422 00:23:52,890 --> 00:23:57,420 became, for a long time the de facto choice and I guess even to 423 00:23:57,420 --> 00:24:01,740 this day, it's still kind of is despite it, there being you 424 00:24:01,740 --> 00:24:06,180 know, hundreds of other options out there. It's just so so 425 00:24:06,180 --> 00:24:06,960 ubiquitous, 426 00:24:07,020 --> 00:24:08,880 John Osborne: it's a good point because I think if you look at 427 00:24:08,880 --> 00:24:14,130 Jenkins is there's there's there's a lot to be said about 428 00:24:14,130 --> 00:24:17,130 it but I think similar to home and you make a good point is 429 00:24:17,130 --> 00:24:20,310 that it's it's the community of plugins and things that are 430 00:24:20,310 --> 00:24:23,400 already out there I can go on to Jenkins and build something out 431 00:24:23,400 --> 00:24:26,100 in an hour that plugs to my Slack channel and does all these 432 00:24:26,100 --> 00:24:30,240 things because those plugins exist and they're easy to 433 00:24:30,240 --> 00:24:33,360 integrate and so it's definitely not a what I would consider like 434 00:24:33,360 --> 00:24:38,220 a cloud native build system but it there's so much there that 435 00:24:38,220 --> 00:24:41,190 you can reuse out of the box that it we don't prefer that you 436 00:24:41,190 --> 00:24:44,550 use at Red Hat we don't prefer that use Jenkins but we get it 437 00:24:44,550 --> 00:24:47,400 if if if that's where you end up landing 438 00:24:47,460 --> 00:24:50,610 Jamie Duncan: Yeah, the copy paste ability of Jenkins once 439 00:24:50,610 --> 00:24:54,930 you have something built, you're able to duplicate it quickly and 440 00:24:54,990 --> 00:24:58,920 enhance because what you find in a because it does very much 441 00:24:58,920 --> 00:25:02,400 mimic your communication And patterns, most of your 442 00:25:02,400 --> 00:25:05,970 applications deploy in a similar way. Once you get the first one 443 00:25:05,970 --> 00:25:08,340 figured out, you know the first pill cost a million dollars a 444 00:25:08,340 --> 00:25:13,170 second pill cost a nickel. Once you get that first pipeline 445 00:25:13,170 --> 00:25:16,590 built out, you can reference it and just make the needed 446 00:25:16,590 --> 00:25:19,530 modifications to build out a pretty mature Jenkins 447 00:25:19,530 --> 00:25:23,880 environment. I deployed Jenkins Actually, this morning, just to 448 00:25:23,880 --> 00:25:25,950 look because I hadn't looked at Jenkins in a long time I 449 00:25:25,980 --> 00:25:30,000 deployed it into a Kubernetes cluster. And the process was 450 00:25:30,000 --> 00:25:34,710 pretty good. You know, so I took on, I said, you know, when we 451 00:25:34,710 --> 00:25:38,280 were planning this first episode out, I said, I would deploy 452 00:25:38,280 --> 00:25:41,460 Jenkins and Spinnaker, and then add a Jenkins X to it this 453 00:25:41,460 --> 00:25:45,690 morning when I was when I was tinkering around. Have those 454 00:25:45,870 --> 00:25:49,950 three? Jenkins is the most straightforward thing to deploy. 455 00:25:50,220 --> 00:25:52,560 Was it a Helm chart? Yeah, of course, it was. 456 00:25:52,650 --> 00:25:53,250 Tariq Islam: I knew it. 457 00:25:54,810 --> 00:25:57,840 John Osborne: Jenkins feels so 2005. But it does Oh my god, it 458 00:25:57,840 --> 00:25:57,990 is. 459 00:26:00,150 --> 00:26:03,660 Jamie Duncan: Like, it's, I'm stunned it how like, it's soon 460 00:26:03,690 --> 00:26:06,300 Alright, so to deploy Jenkins, your there's a little bit of 461 00:26:06,300 --> 00:26:09,840 ammo, it needs a service account inside the coop cluster. You 462 00:26:09,840 --> 00:26:12,540 create a namespace, create a service account, run a Helm 463 00:26:12,540 --> 00:26:15,840 chart, and your cluster is up and running takes three minutes, 464 00:26:15,900 --> 00:26:21,600 I logged into that thing. And I was immediately in 2009. And the 465 00:26:21,600 --> 00:26:24,930 only differences in the UI that I can ascertain after like a 10 466 00:26:24,930 --> 00:26:30,900 minute analysis was they've made the font smaller. And the 467 00:26:30,900 --> 00:26:35,130 graphics were updated about six like they it's like they took 468 00:26:35,130 --> 00:26:39,000 the kitchen from 1998. And they painted the cabinets white and 469 00:26:39,000 --> 00:26:41,850 put down new countertops. But it's the exact same kitchen 470 00:26:41,970 --> 00:26:43,020 Tariq Islam: felt like coming home. 471 00:26:43,410 --> 00:26:47,040 Jamie Duncan: Oh, like, like coming in back to the you know 472 00:26:47,040 --> 00:26:53,040 the step dad that you never really liked them. But they're 473 00:26:53,100 --> 00:26:56,160 very straightforward to deploy into a Kubernetes cluster and 474 00:26:56,160 --> 00:26:59,760 took advantage. Like once it had that service account. It did 475 00:26:59,760 --> 00:27:03,180 populate stuff. So like it was running builds inside the coop 476 00:27:03,180 --> 00:27:05,850 cluster, it was doing everything pretty effectively, pretty 477 00:27:05,850 --> 00:27:10,200 quickly. But I'm just like it was 12 years ago, when you click 478 00:27:10,200 --> 00:27:13,170 new pipeline, you know, no matter which option you pick 479 00:27:13,170 --> 00:27:17,790 whether it was one of the pre canned ones, or the DIY one. It 480 00:27:17,790 --> 00:27:22,110 was 74 radio buttons, 92, checkboxes, and every one of 481 00:27:22,110 --> 00:27:25,230 them brought up for other questions you had to answer. And 482 00:27:25,230 --> 00:27:28,380 at the bottom was a text area that I wrote my shell script in? 483 00:27:28,920 --> 00:27:32,580 Yeah, it is. And it was I was like, Oh, my it like the 484 00:27:32,580 --> 00:27:35,220 workflow. And I guess at that point, because it is so 485 00:27:35,220 --> 00:27:39,030 ubiquitous, it's rate of evolution has to be incredibly 486 00:27:39,030 --> 00:27:43,410 slow. It's the generic general purpose, right? The old server, 487 00:27:43,410 --> 00:27:45,930 everyone uses it. So everyone's holding on to the way they're 488 00:27:45,930 --> 00:27:49,380 currently using it. I feel really sorry for cloudbees you 489 00:27:49,380 --> 00:27:51,000 know, the people that do it professionally? 490 00:27:51,810 --> 00:27:53,460 John Osborne: Yeah, there's something to that too. Because 491 00:27:53,460 --> 00:27:57,600 if you look at the a lot of the way, you know, in practice, at 492 00:27:57,600 --> 00:28:00,330 least what I've seen, tell me if you've seen this, but when 493 00:28:00,360 --> 00:28:03,420 people get hired as a quote unquote DevOps engineer, what 494 00:28:03,420 --> 00:28:05,820 they end up doing is just writing a lot of glucose, and 495 00:28:05,820 --> 00:28:09,960 that glucose ends up often living as custom shell scripts 496 00:28:09,990 --> 00:28:14,160 inside of Jenkins. Yeah, and yeah, that's a hard place to 497 00:28:14,160 --> 00:28:14,910 live in, I think, 498 00:28:15,000 --> 00:28:17,730 Jamie Duncan: yeah. It reminded It reminds me a lot of like, the 499 00:28:17,730 --> 00:28:22,950 way pivotal stuff works. reminds me a lot of the way I did jobs 500 00:28:22,980 --> 00:28:24,060 1213 years ago, 501 00:28:24,300 --> 00:28:27,120 Tariq Islam: I will say that, you know, I'm a fan of knocking 502 00:28:27,120 --> 00:28:32,490 on Jenkins, but for, for the, for the for the listeners. And I 503 00:28:32,490 --> 00:28:36,000 still think that if if it does end up being Jenkins that you do 504 00:28:36,000 --> 00:28:38,940 go with as your you're kind of getting started stepping stone, 505 00:28:39,150 --> 00:28:42,060 so be it right. And if that's what that's what you're used to 506 00:28:42,060 --> 00:28:45,060 get good with ci CD and automation and getting, getting 507 00:28:45,060 --> 00:28:48,930 to a better place. Go for it. Now, that said, You are 508 00:28:48,960 --> 00:28:52,260 absolutely going to hit, you know, major issues with respect 509 00:28:52,260 --> 00:28:57,090 to scaling. You know, Jenkins is not exactly friendly to 510 00:28:57,090 --> 00:28:59,850 resources. But that's something that that's something 511 00:28:59,849 --> 00:29:02,009 John Osborne: you might care about day one. 512 00:29:02,099 --> 00:29:05,039 Jamie Duncan: Yeah. And like you said, starting at one user, it 513 00:29:05,039 --> 00:29:09,839 works just fine. Mm hmm. Yeah, yeah. It really scaling up in 514 00:29:09,839 --> 00:29:14,399 102,000. I mean, let's all you know, I freely admit that the 515 00:29:14,399 --> 00:29:18,689 modern internet owns lens, a large portion of its existence 516 00:29:18,689 --> 00:29:22,499 to Jenkins, is you can't get that much software deployed 517 00:29:22,499 --> 00:29:27,209 manually. We would still be working on your stuff from seven 518 00:29:27,209 --> 00:29:30,659 years ago. We all have benefited from Jenkins globally as a 519 00:29:30,659 --> 00:29:34,859 product that said, it's like Domino's Pizza. It is not the 520 00:29:34,859 --> 00:29:38,159 best pizza in the town you live in, but it's going to show up 521 00:29:38,159 --> 00:29:41,249 pretty consistently better than they're going. Right, exactly. 522 00:29:41,819 --> 00:29:43,739 They're going to be there in a half hour and you're going to 523 00:29:43,739 --> 00:29:48,599 eat it. Yeah. It's not the best pizza on the planet. You can get 524 00:29:48,599 --> 00:29:52,229 a better experience. But, but you're going to pay for it and 525 00:29:52,229 --> 00:29:55,259 you're going to have to jump through some hoops to get it. So 526 00:29:55,529 --> 00:29:58,949 one of those better experiences. So what Jenkins ended up doing a 527 00:29:58,949 --> 00:30:03,059 few years ago, and I don't know exactly when. But Jenkins then 528 00:30:03,659 --> 00:30:07,319 is, is there and it has a sort of glacial evolutionary path, 529 00:30:07,379 --> 00:30:09,419 you know, because everyone's using it. No one wants it to 530 00:30:09,419 --> 00:30:12,809 change. So what they did was make a thing called Jenkins x. 531 00:30:13,349 --> 00:30:16,079 It was back when I think putting x in front of at the end of 532 00:30:16,079 --> 00:30:18,479 things was cool. That was actually a team that left Red 533 00:30:18,479 --> 00:30:22,259 Hat to, to go build that. And if you remember, was it. Yep. It 534 00:30:22,259 --> 00:30:23,789 was. Oh, yeah. Yeah, not. 535 00:30:24,720 --> 00:30:28,980 John Osborne: So. And it wasn't the same Jenkins team. It was a 536 00:30:28,980 --> 00:30:31,950 bunch of people that were really strong and on messaging and kind 537 00:30:31,950 --> 00:30:35,280 of middleware. Some of the people that worked on are built 538 00:30:35,280 --> 00:30:39,750 Apache camel, those types of people, and they left and did 539 00:30:39,750 --> 00:30:42,240 their own thing. And they came up with Jenkins x. 540 00:30:42,599 --> 00:30:45,899 Jamie Duncan: Yep. So Jenkins X has a great UI for Kubernetes 541 00:30:45,929 --> 00:30:49,649 ties itself into octant, which is a really solid user 542 00:30:49,649 --> 00:30:53,969 experience. There's an ops persona panel for for Jenkins x, 543 00:30:53,969 --> 00:30:57,179 there's a dev persona panel for Jenkins x, really slick, really 544 00:30:57,179 --> 00:31:01,799 modern, get a lot of work done really efficiently. So I think 545 00:31:01,799 --> 00:31:04,889 under the covers, it's using a lot of the same state engine 546 00:31:04,889 --> 00:31:09,119 tooling. It's a lot of the same engine of Jenkins, but it's 547 00:31:09,179 --> 00:31:11,609 presented in a more modern fashion. And it's got more 548 00:31:11,609 --> 00:31:14,039 modern hooks, like there's a more modern API that you can 549 00:31:14,039 --> 00:31:18,119 interact with, and really deal with the tooling much more in a 550 00:31:18,119 --> 00:31:21,929 much more modern cloud native way. deploying it to some of the 551 00:31:21,929 --> 00:31:25,619 silly stuff I've ever seen in my life, I cannot figure out the 552 00:31:25,619 --> 00:31:30,149 lot. The logic of how they came up with the deployment tooling 553 00:31:30,179 --> 00:31:35,669 for Jenkins x. It starts out really cool. It starts out easy. 554 00:31:36,209 --> 00:31:40,559 Like it's like 10 lines of JSON, that you drop into a terraform 555 00:31:40,589 --> 00:31:45,089 template, and you run terraform in it and terraform apply. And 556 00:31:45,089 --> 00:31:49,409 then it checks out a repository and it does Oh, no, then it does 557 00:31:49,409 --> 00:31:52,829 some stuff. So it actually deploys a Kubernetes cluster on 558 00:31:52,829 --> 00:31:57,029 your cloud provider. So like you tell them in that in that JSON, 559 00:31:57,029 --> 00:32:00,719 you tell it what cloud project you're using in different ones, 560 00:32:00,719 --> 00:32:04,169 like AWS looks a lot different than Google Cloud, all that. But 561 00:32:04,169 --> 00:32:06,449 it stands up a Kubernetes cluster, and in that Kubernetes 562 00:32:06,449 --> 00:32:10,229 clusters running just about nothing, then you have to go 563 00:32:10,259 --> 00:32:16,439 download another binary, called j X on your local machine, run a 564 00:32:16,439 --> 00:32:20,099 couple of more commands, one of which checks out a repository 565 00:32:20,099 --> 00:32:25,019 from GitHub. And then that generates a yamo file that you 566 00:32:25,019 --> 00:32:28,319 run another command on that thing goes and sticks the right 567 00:32:28,319 --> 00:32:32,249 things in the Kubernetes cluster that you deployed 20 minutes 568 00:32:32,249 --> 00:32:32,849 previously. 569 00:32:33,869 --> 00:32:37,829 John Osborne: I say felted, just when you're explaining, Yeah, 570 00:32:37,860 --> 00:32:38,790 Jamie Duncan: dude, I and 571 00:32:39,030 --> 00:32:41,460 John Osborne: I looked at Jenkins x when I first started 572 00:32:41,460 --> 00:32:42,840 that company. So I was interested in what they were 573 00:32:42,840 --> 00:32:45,030 doing. But I haven't really followed it much the last couple 574 00:32:45,030 --> 00:32:45,480 years. 575 00:32:46,080 --> 00:32:48,780 Jamie Duncan: I don't understand how they got there. And the fun 576 00:32:48,780 --> 00:32:51,990 thing is I pull down the latest version of everything. There is 577 00:32:51,990 --> 00:32:55,050 some hardware inside some of that. And it's all terraform 578 00:32:55,050 --> 00:32:57,990 under the covers the deployment tooling inside some of the 579 00:32:57,990 --> 00:33:01,200 terraform, or inside some of the file somewhere inside this JSX 580 00:33:01,200 --> 00:33:05,430 binary, it is hard coded to look to check out the master from 581 00:33:05,430 --> 00:33:08,160 when it checks out this GitHub repo, and the master branch 582 00:33:08,160 --> 00:33:10,290 doesn't exist anymore. And GitHub because they renamed it 583 00:33:10,290 --> 00:33:14,910 to Maine, but the installer hasn't been updated. And I 584 00:33:14,910 --> 00:33:18,840 pulled down the latest release binary to try it out for like 585 00:33:18,840 --> 00:33:22,080 three different versions. So the installer for Jenkins x is 586 00:33:22,080 --> 00:33:23,400 effectively broken right now. 587 00:33:24,030 --> 00:33:28,530 Tariq Islam: So this is where managing your own 588 00:33:28,710 --> 00:33:33,030 infrastructure. And the type of infrastructure you choose. As 589 00:33:33,030 --> 00:33:38,370 soon as you get beyond a small team really starts to matter. I 590 00:33:38,369 --> 00:33:41,849 Jamie Duncan: think so yeah. So when we were sketching this, 591 00:33:41,879 --> 00:33:48,779 this episode out, we broke down three generic categories of CD 592 00:33:48,779 --> 00:33:53,399 backbone CD platforms, Jenkins and Spinnaker would be the 593 00:33:53,399 --> 00:33:55,529 number two one, but we're not talking to so Jenkins and 594 00:33:55,529 --> 00:33:58,829 chicken neck Spinnaker, and I'm sure there are others are what 595 00:33:58,829 --> 00:34:02,189 you would call traditional. And those are the you know, I'm 596 00:34:02,189 --> 00:34:04,019 going to deploy a thing and whether it's deployed into 597 00:34:04,019 --> 00:34:08,129 Kubernetes or deployed into a VM or wherever. It's a traditional 598 00:34:08,129 --> 00:34:11,009 application, and I get a traditional experience from it. 599 00:34:11,219 --> 00:34:16,349 The second type C FCD. backbone is one that's a little more 600 00:34:16,349 --> 00:34:20,909 cloud native. And is tark. is do you want to get into that a 601 00:34:20,909 --> 00:34:21,479 little bit. 602 00:34:21,690 --> 00:34:25,020 Tariq Islam: Yes, so the cloud native ones, some some examples 603 00:34:25,020 --> 00:34:29,340 here. But this is by no means an exhaustive list. In fact, this 604 00:34:29,340 --> 00:34:33,570 is probably 1% of everything that's out there, you know, 605 00:34:33,570 --> 00:34:40,680 tools like circle ci, GitHub actions, git lab, ci, Travis CI, 606 00:34:40,920 --> 00:34:45,750 I think of those really comes out to, you know, GitHub and get 607 00:34:45,750 --> 00:34:48,720 lab being the more popular ones. I've seen more folks, I've 608 00:34:48,720 --> 00:34:51,750 actually seen quite a lot of Git lab. 609 00:34:52,199 --> 00:34:54,779 Jamie Duncan: Yeah, it's really good. I used it a lot. We used 610 00:34:54,779 --> 00:35:01,079 it a lot at VMware. really effective. Travis and circle I 611 00:35:01,079 --> 00:35:03,929 think they were they were first to market in this space. And I 612 00:35:03,929 --> 00:35:08,519 don't know who be who. But Travis and circle, were filling 613 00:35:08,519 --> 00:35:11,789 the gap before get lab and GitHub had their own integrated 614 00:35:11,789 --> 00:35:12,119 CD 615 00:35:12,120 --> 00:35:15,120 Tariq Islam: platform. For my own edification, how are we 616 00:35:15,150 --> 00:35:18,060 differentiating between traditional and cloud native? Is 617 00:35:18,060 --> 00:35:22,530 it just how many years old they are? Or deployment model? Maybe 618 00:35:22,530 --> 00:35:26,640 both? What makes a thing cloud native in this context? 619 00:35:27,180 --> 00:35:29,610 John Osborne: I mean, I think, you know, if you look at tecton, 620 00:35:29,610 --> 00:35:34,110 or some of the newer ways to build pipelines, it's that they 621 00:35:34,110 --> 00:35:39,420 can one be event driven, in the sense that, you know, something 622 00:35:39,420 --> 00:35:44,520 can trigger the pipeline to start or even instantiate from 623 00:35:44,520 --> 00:35:48,690 scratch, right? And then, you know, does it run well, in 624 00:35:48,690 --> 00:35:54,870 containers? Can you scale up scale down? Very easily? You 625 00:35:54,870 --> 00:35:57,600 know, is it a lot of times now, we're also talking about is it 626 00:35:57,600 --> 00:36:00,870 declarative in the sense that you set what the desired state 627 00:36:00,870 --> 00:36:05,850 is going to be? And not just the steps to get there an imperative 628 00:36:05,850 --> 00:36:10,290 model? But it's, you know, is it going to run on Kubernetes? 629 00:36:10,290 --> 00:36:13,860 Well, is it going to run as a container? Well, those are the 630 00:36:13,860 --> 00:36:16,530 types of things that I that I normally think about, I will say 631 00:36:16,530 --> 00:36:21,630 to like, before, before I forget, you know, I think when 632 00:36:21,630 --> 00:36:26,970 we talk about building, in a pipeline, any sort of scale like 633 00:36:27,000 --> 00:36:29,340 and I would even say, once you get to like 10 nodes, if you're 634 00:36:29,340 --> 00:36:33,300 building in your cluster, you probably want a dedicated 635 00:36:33,300 --> 00:36:37,200 machine just for the builds, because that he comes so spiky, 636 00:36:37,920 --> 00:36:44,700 that it really can destroy some of your, your, your node 637 00:36:45,000 --> 00:36:48,510 resource usage, you know, things can spike up really high spike 638 00:36:48,510 --> 00:36:50,970 back down, especially if you're doing Java, you know, start 639 00:36:50,970 --> 00:36:55,770 doing Maven builds or even node j s, you know, Mo the NPM stuff. 640 00:36:57,240 --> 00:37:00,360 You know, even go right, because it's it's statically compiled. 641 00:37:00,360 --> 00:37:04,770 So the builds tend to be pretty compute resource intensive 642 00:37:04,770 --> 00:37:07,350 first, but obviously, the runtime is very good, right. So 643 00:37:08,850 --> 00:37:11,490 you know, you probably at any sort of even medium scale, you 644 00:37:11,490 --> 00:37:15,510 probably want to look to start to have dedicated machines that 645 00:37:15,510 --> 00:37:17,340 just do builds. This sounds 646 00:37:17,340 --> 00:37:19,680 Tariq Islam: a lot like the problems that we've seen 647 00:37:19,680 --> 00:37:24,270 organizations run into with, with Jenkins, right, is resource 648 00:37:24,270 --> 00:37:28,410 contention spike Enos having to have dedicated infrastructure, I 649 00:37:28,410 --> 00:37:32,340 think, you know, I don't know, maybe maybe cloud native is not 650 00:37:32,340 --> 00:37:35,130 the most accurate term for this category, maybe it's the fact 651 00:37:35,130 --> 00:37:40,830 that the design principles, the user experience, in these kind 652 00:37:40,830 --> 00:37:48,720 of Next Gen. backbone, backbones are, are just better, you know, 653 00:37:48,720 --> 00:37:49,770 a little bit more refined. 654 00:37:50,490 --> 00:37:53,160 Jamie Duncan: And I'm thinking through, because I think I was 655 00:37:53,160 --> 00:37:55,920 the one that wrote cloud native in the outline, so I'll take the 656 00:37:55,920 --> 00:37:56,550 hit for that one. 657 00:37:57,720 --> 00:37:58,770 John Osborne: Actually, I didn't even see that. 658 00:38:01,800 --> 00:38:04,770 Jamie Duncan: But I'm gonna hold on to I'm not gonna let go of it 659 00:38:04,770 --> 00:38:09,750 just yet. Because john tark, I think you're 100% on point is 660 00:38:09,780 --> 00:38:14,640 that they are an evolution and how it works. The ones that we 661 00:38:14,670 --> 00:38:17,520 the ones that we've taught that we've sort of enumerated circle 662 00:38:17,520 --> 00:38:21,330 and Travis and get lab and get hubs integrated ones, and there 663 00:38:21,330 --> 00:38:25,110 are a ton more and a ton of them are very platform specific, like 664 00:38:25,110 --> 00:38:30,390 natla, fly and a bunch of others. Even if you're deploying 665 00:38:30,390 --> 00:38:34,890 get lab on prem, and you do have to account for those build 666 00:38:34,890 --> 00:38:41,040 resources. When I'm hooking Mike a Kubernetes, cluster into those 667 00:38:41,040 --> 00:38:47,250 build resources, it takes me about 30 seconds. And the entire 668 00:38:47,250 --> 00:38:51,600 experience is driven through the Git lab UI. And I effectively 669 00:38:51,600 --> 00:38:55,920 just give Git lab permissions to go deploy workloads on my 670 00:38:55,920 --> 00:38:59,700 cluster. And I can drop it into a namespace with quotas and 671 00:38:59,700 --> 00:39:04,650 handle those resources that way. So we're Git lab, because it has 672 00:39:04,650 --> 00:39:09,360 an on prem version is almost a hybrid between a traditional and 673 00:39:09,360 --> 00:39:13,650 a cloud native. You do account for that to some degree with 674 00:39:13,650 --> 00:39:16,440 with Jenkins. And I've actually ran in or give with Git lab. And 675 00:39:16,440 --> 00:39:19,320 I've actually ran into that where, you know, my build server 676 00:39:19,470 --> 00:39:22,830 ran out of resources a lot. It was some projects I worked on at 677 00:39:22,830 --> 00:39:28,140 VMware. But with like GitHub, and Travis and circle, they're 678 00:39:28,140 --> 00:39:32,070 all as a service. So you don't have to provide any of you don't 679 00:39:32,070 --> 00:39:36,240 have to account for any of it. And the worst case scenario is 680 00:39:36,240 --> 00:39:38,580 that you're waiting a little while longer on your build to 681 00:39:38,580 --> 00:39:41,580 run if it's eating up a lot of stuff, or if the run queue is 682 00:39:41,580 --> 00:39:45,870 just long with GitHub actions or, you know, in the Travis CI, 683 00:39:46,080 --> 00:39:48,870 Travis is handling all that for you or circle ci is handling all 684 00:39:48,870 --> 00:39:49,470 that for you. 685 00:39:51,060 --> 00:39:52,740 John Osborne: They've done a really good job. Some of those 686 00:39:52,740 --> 00:39:54,090 tools were Oh man, 687 00:39:54,120 --> 00:39:54,810 Jamie Duncan: they are awesome. 688 00:39:54,960 --> 00:39:59,430 John Osborne: They used to be a Git repo and now their ci CD 689 00:39:59,430 --> 00:40:05,820 workflows. thinly disguised as a Git repo a Git repo. Yeah. But 690 00:40:05,820 --> 00:40:07,140 they've done a good job for sure. 691 00:40:07,200 --> 00:40:10,110 Jamie Duncan: Yeah. And the cool thing about all of these is the 692 00:40:10,110 --> 00:40:14,190 way they present themselves to the user. And it's kind of the 693 00:40:14,190 --> 00:40:18,450 reason I'm holding on to cloud native as them as a moniker for 694 00:40:18,450 --> 00:40:22,470 them. Because all of them have the same workflow. Whereas 695 00:40:22,470 --> 00:40:25,410 Jenkins, you've got to go into the Jenkins UI to find these 696 00:40:25,410 --> 00:40:28,320 things, you know, define a pipeline to get all this stuff 697 00:40:28,320 --> 00:40:31,110 hammered out all these buttons and checkboxes and text fields. 698 00:40:32,520 --> 00:40:35,070 With all of the cloud native tools that we've talked that we 699 00:40:35,070 --> 00:40:39,240 just mentioned, all of them are a single YAML file at the base 700 00:40:39,240 --> 00:40:44,940 of your repository. And that your entire experience. That 701 00:40:44,940 --> 00:40:49,110 means that it's massively powerful. It is Yep. And that's 702 00:40:49,110 --> 00:40:51,720 what that's why I'm holding on to cloud native, because if I 703 00:40:51,720 --> 00:40:57,630 drop a YAML, dot, Travis, ci dot YAML, at the top of my repo, and 704 00:40:57,630 --> 00:41:00,330 I tell Travis that you have permissions to talk to this 705 00:41:00,330 --> 00:41:02,580 repo, and I'm done. 706 00:41:03,030 --> 00:41:05,940 Tariq Islam: So this begs the question, then is it worth, you 707 00:41:05,940 --> 00:41:08,430 know, whether it's the organization that's still 708 00:41:08,430 --> 00:41:12,360 emailing their zip files? Or maybe it's another organization 709 00:41:12,360 --> 00:41:15,720 that's slightly more mature than that, or retelling them? Or we 710 00:41:15,720 --> 00:41:17,760 issue like a peanut butter spread? Going back with peanut 711 00:41:17,760 --> 00:41:18,270 butter here? 712 00:41:18,900 --> 00:41:20,460 Jamie Duncan: Are we just telling everyone always goes 713 00:41:20,460 --> 00:41:21,000 back to penguin? 714 00:41:21,780 --> 00:41:24,660 Tariq Islam: are we telling folks, you know, just go 715 00:41:24,660 --> 00:41:29,910 straight for the, the what we're calling cloud native, like the 716 00:41:29,910 --> 00:41:33,570 circle ci, and the Travis CI is and get labs of the world? And 717 00:41:33,570 --> 00:41:37,530 skipping over, over Jenkins? And just maybe, are we saying that 718 00:41:37,530 --> 00:41:43,770 Jenkins is just legacy enough? That you can you can you can 719 00:41:43,770 --> 00:41:46,050 forget about it and move into these newer tools? 720 00:41:47,010 --> 00:41:49,590 John Osborne: I think it depends. Because if you look at, 721 00:41:49,650 --> 00:41:52,920 you know, cloud native stuff, right, you know, you has all its 722 00:41:52,920 --> 00:41:55,590 dependencies with it kind of handles restarts, well, you 723 00:41:55,590 --> 00:41:58,050 know, it can run on containers, Kubernetes, it's highly 724 00:41:58,050 --> 00:42:02,490 available, all that stuff. You know, that requires in some 725 00:42:02,490 --> 00:42:05,880 cases, you know, so you're using tecton, as an example, right. So 726 00:42:07,050 --> 00:42:09,480 you might have to write containers with automation in it 727 00:42:09,480 --> 00:42:12,720 to make some of this stuff happen that you might have a 728 00:42:12,720 --> 00:42:15,630 plugin out of the box with Jenkins. So it's, it is kind of 729 00:42:15,630 --> 00:42:18,780 a your mileage may vary, depending on your use case, but 730 00:42:19,380 --> 00:42:23,310 and not every org has people just sitting around that are 731 00:42:23,340 --> 00:42:27,210 writing Glue Code in containers, to do all this stuff. So it's 732 00:42:27,210 --> 00:42:30,570 kind of a, you know, you have to make your own decision. But I do 733 00:42:30,570 --> 00:42:32,610 think that there's a certain amount of technical debt you 734 00:42:32,610 --> 00:42:37,350 assume by by doing Greenfield Jenkins deployments at this 735 00:42:37,350 --> 00:42:37,680 point. 736 00:42:38,820 --> 00:42:41,970 Jamie Duncan: Yeah. And I think at least for me, I see them as 737 00:42:41,970 --> 00:42:49,020 a, it's a case coverage problem is circle and Travis and GitHub 738 00:42:49,020 --> 00:42:54,660 and get lab, they all have some limitations. If you're going to 739 00:42:54,660 --> 00:42:58,260 use Travis or circle, or just about any of them, I think 740 00:42:59,010 --> 00:43:03,000 you're going to be in the limitations aren't great. But 741 00:43:03,000 --> 00:43:06,720 you're going to have to use some of their tooling. And they're 742 00:43:06,720 --> 00:43:10,020 going to make some assumptions about what you're doing. And the 743 00:43:10,020 --> 00:43:12,660 steps are going to be relative or somewhat limited compared to 744 00:43:12,660 --> 00:43:15,360 what you can do in Travis, or what you can do in Jenkins, 745 00:43:15,360 --> 00:43:19,140 where you know, the world is your is your checkbox. If you 746 00:43:19,140 --> 00:43:23,730 find yourself straining against those limitations be because of 747 00:43:23,730 --> 00:43:25,890 the way you've designed your application, or because of the 748 00:43:25,890 --> 00:43:30,060 limitations or security needs you've got. You always have 749 00:43:30,060 --> 00:43:34,140 Jenkins available as this infinitely customizable platform 750 00:43:34,140 --> 00:43:37,440 that's going to get you where you need to go. But for that 751 00:43:37,500 --> 00:43:41,340 problem, I don't know if it's 85, or 97 and a half. But for a 752 00:43:41,340 --> 00:43:44,100 very high percentage of modern applications being written 753 00:43:44,100 --> 00:43:48,660 today. These cloud native CD tools, fit your use case very 754 00:43:48,660 --> 00:43:53,520 comfortably, and are a much lower barrier to entry into this 755 00:43:53,520 --> 00:43:53,940 arena. 756 00:43:54,420 --> 00:43:57,150 Tariq Islam: And I will say that of these, or at least of the 757 00:43:57,150 --> 00:43:59,670 ones that I've seen, I think I think Git lab is probably the 758 00:43:59,670 --> 00:44:05,460 furthest along as far as bringing together feature set 759 00:44:05,580 --> 00:44:12,990 ease of use deployment pattern. You know it? Again, to your 760 00:44:12,990 --> 00:44:16,800 point, Jamie, it's not it's not 100%. As far as customizability 761 00:44:16,800 --> 00:44:20,430 the way Jenkins is, but I think I think for the vast majority of 762 00:44:20,430 --> 00:44:22,920 organizations, they've I'm calling them out because you 763 00:44:22,920 --> 00:44:24,030 know, they've done great work. 764 00:44:24,180 --> 00:44:27,420 Jamie Duncan: Yeah, I agree. I like I said, I use it pretty 765 00:44:27,420 --> 00:44:30,600 extensively at VMware, and have used it on and off before and 766 00:44:30,600 --> 00:44:34,740 after. And I'd agree, yeah, they are the most mature of all the 767 00:44:34,740 --> 00:44:40,920 cloud native ci, CD tooling. And it is very easy to use. 768 00:44:41,310 --> 00:44:43,380 Tariq Islam: So let's talk about this this last category that 769 00:44:43,380 --> 00:44:48,780 we've got here, managed, right and managed by managed and the 770 00:44:48,780 --> 00:44:52,110 most sensitive to this term. Because I think there's a lot of 771 00:44:52,110 --> 00:44:56,640 confusion around it. Anything for me, and we I don't I don't 772 00:44:56,640 --> 00:45:00,570 want to make this a rat hole. But when I say manage I'm 773 00:45:00,570 --> 00:45:06,090 talking about, you are not deploying this infrastructure or 774 00:45:06,090 --> 00:45:11,220 this tool in any way, shape or form where you're having to 775 00:45:11,280 --> 00:45:14,550 manage it. Right? There's no such thing as day to operations 776 00:45:14,550 --> 00:45:23,580 here. This is a purely Oh serverless for form of, of ci CD 777 00:45:23,610 --> 00:45:27,990 infrastructure, where you just kind of click on a button. And 778 00:45:28,680 --> 00:45:31,260 and you get a thing. Yeah. 779 00:45:32,100 --> 00:45:35,130 Jamie Duncan: Right. Yeah, I see them is in the what we're trying 780 00:45:35,130 --> 00:45:37,440 to talk the examples of what we're talking about her, like 781 00:45:37,470 --> 00:45:42,330 AWS code pipeline, Google's cloud build, we couldn't come up 782 00:45:42,330 --> 00:45:44,910 with azures, I think they've called it Azure DevOps. Yeah, 783 00:45:44,910 --> 00:45:46,320 it's their builds that 784 00:45:46,469 --> 00:45:49,469 John Osborne: it's that that the TFS, the Team Foundation Server 785 00:45:49,469 --> 00:45:51,599 stuff in Azure called Azure DevOps. 786 00:45:52,170 --> 00:45:57,390 Jamie Duncan: Yeah. And really what you get as you get a pretty 787 00:45:57,540 --> 00:46:02,520 configurable set of inputs. And you get a completely abstracted 788 00:46:02,520 --> 00:46:06,810 build experience, and you get a less configurable set of 789 00:46:06,810 --> 00:46:10,560 destinations or output of what it will what it will generate 790 00:46:10,560 --> 00:46:14,070 for you and where it will put it. But that's really it, you 791 00:46:14,070 --> 00:46:17,040 feed it your source code, you feed it some parameters, it does 792 00:46:17,040 --> 00:46:21,030 some magic, and it drops the result of that magic in one of a 793 00:46:21,030 --> 00:46:22,140 handful of locations. 794 00:46:23,550 --> 00:46:26,490 Tariq Islam: So I think this one, this category is pretty 795 00:46:26,490 --> 00:46:31,020 interesting, because I feel like it is the most scalable from one 796 00:46:31,020 --> 00:46:35,040 to 100,000, just because of the nature of how its deployed, you 797 00:46:35,040 --> 00:46:37,710 know, at least from the perspective of the user, it's 798 00:46:37,710 --> 00:46:40,320 not something that's deployed. But you know, me as an 799 00:46:40,320 --> 00:46:43,440 individual, I can use Cloud build, for example, or Azure 800 00:46:43,440 --> 00:46:47,700 DevOps for just my own purposes. And then I could bring any of 801 00:46:47,700 --> 00:46:52,140 these into a fully fledged enterprise organization, and use 802 00:46:52,140 --> 00:46:56,850 them at scale as well. Now, you know, you do have to consider 803 00:46:57,480 --> 00:47:01,080 which one of these you're going to pick, because the one that 804 00:47:01,080 --> 00:47:05,040 you pick, is is going to is going to effectively tie you to 805 00:47:05,040 --> 00:47:10,350 a cloud. Right, but, but i think i think i think listeners 806 00:47:10,350 --> 00:47:14,520 should, you know, see the advantages of a managed offering 807 00:47:14,520 --> 00:47:18,930 here, because of how quickly you can get up and running, and how 808 00:47:18,930 --> 00:47:21,990 quickly you're able to scale and bring it to the rest of an 809 00:47:21,990 --> 00:47:27,390 organization. But, again, you know, going back to the point 810 00:47:27,390 --> 00:47:29,730 that Jamie made around technical debt, there's something to 811 00:47:29,730 --> 00:47:30,330 consider. 812 00:47:31,530 --> 00:47:33,750 John Osborne: So for my consistency, what you'll see 813 00:47:33,750 --> 00:47:38,910 from adoption, but in my opinion, like ci CD tools in the 814 00:47:38,910 --> 00:47:42,060 cloud, is managed services have some of the lowest adoption of 815 00:47:42,060 --> 00:47:47,100 all the public cloud provider services. And I don't know, but 816 00:47:47,100 --> 00:47:52,440 I think the reason for that is is, is people have a really hard 817 00:47:52,440 --> 00:47:57,420 time paying to build their code. There's some there must be some 818 00:47:57,420 --> 00:48:01,620 psychology there, I don't know. But pay as you go. And 819 00:48:01,620 --> 00:48:05,400 obviously, you pay with, if you manage anything yourself, you 820 00:48:05,460 --> 00:48:11,100 you pay to, in terms of in terms of all the people that you that 821 00:48:11,100 --> 00:48:14,310 are managing those things, right. But it's not as apparent 822 00:48:14,310 --> 00:48:21,390 as paying for a service. And I see challenges with with people 823 00:48:21,390 --> 00:48:25,380 just mentally getting over that hurdle of paying to compile my 824 00:48:25,380 --> 00:48:29,790 code. Have you seen that or such as something that that, that 825 00:48:29,790 --> 00:48:30,330 I've seen? 826 00:48:32,640 --> 00:48:36,300 Tariq Islam: I've actually not seen that yet. Although, again, 827 00:48:36,300 --> 00:48:42,930 I do, I would have to agree with you that it's, it's not it, 828 00:48:43,050 --> 00:48:49,620 there hasn't been a huge wave of adoption across these tools. But 829 00:48:49,620 --> 00:48:57,030 I can, I can see, I can see why that would be prevalent. It's 830 00:48:57,030 --> 00:49:00,780 one of those. It's one of those, like, you know, I'm going to do 831 00:49:00,780 --> 00:49:05,310 it myself kind of thing. It's, it's this is low value task. And 832 00:49:05,310 --> 00:49:07,920 and that's the that's the perception, right, that the 833 00:49:07,920 --> 00:49:13,200 tasks of building, you know, compiling, building, you know, 834 00:49:13,920 --> 00:49:18,240 all of that, all that sci fi stuff, especially, is deemed low 835 00:49:18,240 --> 00:49:24,480 level. machinery. And, and and it's Yeah, I can, I can totally 836 00:49:24,480 --> 00:49:24,930 get there. 837 00:49:25,950 --> 00:49:29,550 Jamie Duncan: Yeah, my take on it was, I certainly agree that 838 00:49:29,580 --> 00:49:32,400 it's not, you know, people aren't beating down the doors to 839 00:49:32,400 --> 00:49:37,200 Google Cloud to use Cloud build. It's a great tool, and it does 840 00:49:37,200 --> 00:49:40,830 exactly what it needs to do. But it's it's very much a value add 841 00:49:40,830 --> 00:49:44,430 for someone who is sort of all in on AWS cloud, at least in the 842 00:49:44,430 --> 00:49:50,280 public cloud space. I assumed that slower curve of adoption 843 00:49:50,280 --> 00:49:53,970 was because people have relatively mature CD pipelines 844 00:49:53,970 --> 00:49:55,800 because like we said, the beginning people have been doing 845 00:49:55,800 --> 00:49:59,430 this for a long time. So if they've got something and it's 846 00:49:59,430 --> 00:50:02,580 not broke, They don't see it as debt, they see it as providing 847 00:50:02,580 --> 00:50:06,060 value. Why would they read? You know, why would they build 848 00:50:06,060 --> 00:50:10,380 cycles to change it? And as you were talking, I started 849 00:50:10,380 --> 00:50:16,770 wondering if that is if it's that as much, or if it's is, as 850 00:50:16,770 --> 00:50:20,490 we were looking at these and talking about them, you know, as 851 00:50:20,490 --> 00:50:23,580 you go from these traditional platforms like Jenkins, to these 852 00:50:23,580 --> 00:50:27,450 cloud native platforms like Git lab, to these managed platforms, 853 00:50:27,480 --> 00:50:34,860 like like AWS code pipeline, each step along the way, you 854 00:50:34,860 --> 00:50:40,380 have to have good, better hygiene. Like you can do some 855 00:50:40,380 --> 00:50:43,350 crazy woolly stuff inside Jenkins and get your application 856 00:50:43,350 --> 00:50:43,800 to build. 857 00:50:44,400 --> 00:50:46,260 John Osborne: Yeah, automation is hard. I mean, 858 00:50:46,440 --> 00:50:48,360 Jamie Duncan: yeah, exactly. Like we said at the beginning, 859 00:50:48,870 --> 00:50:50,910 but and there's lots of duct tape is strong. 860 00:50:51,180 --> 00:50:54,870 John Osborne: Yeah. Duct Tape survival solutions, sometimes 861 00:50:54,870 --> 00:50:55,710 Exactly. Oh, yeah. 862 00:50:55,709 --> 00:50:58,169 Jamie Duncan: And Jenkins is really good duct tape. It's the 863 00:50:58,169 --> 00:51:03,089 duct tape of the internet. It is needed. Yeah, the cloud native 864 00:51:03,089 --> 00:51:06,359 tools, you can use a little less duct tape, you have to have a 865 00:51:06,359 --> 00:51:09,569 little better hygiene, there are some more limitations. And the 866 00:51:09,569 --> 00:51:12,359 Manage platforms, there are even more limitations. Like I said, 867 00:51:12,359 --> 00:51:14,849 you have a more limited set of inputs, you have a more limited 868 00:51:14,849 --> 00:51:19,019 set of outputs. And you don't have a lot of control of what 869 00:51:19,019 --> 00:51:22,979 happens in the middle. So I wonder if it's because they're 870 00:51:22,979 --> 00:51:26,699 just they have something they're happy with? Or it's because 871 00:51:26,699 --> 00:51:29,489 they've got their house of cards duct taped up enough that they 872 00:51:29,489 --> 00:51:34,829 don't want to touch it. That is that sort of limiting some of 873 00:51:34,829 --> 00:51:38,489 that, that that adoption of these managed platforms? I don't 874 00:51:38,489 --> 00:51:41,189 know, it's kind of an open ended question. I'm thinking about it, 875 00:51:41,249 --> 00:51:45,149 you know, even as I say it out loud. But I'm, we're running a 876 00:51:45,149 --> 00:51:48,659 little long here. Yeah, talking about this. So john, do you want 877 00:51:48,659 --> 00:51:51,449 to give us give the skinny on what we're going to do for the 878 00:51:51,449 --> 00:51:55,439 next five or six episodes, yes, and throw this idea out for 879 00:51:55,650 --> 00:51:57,420 John Osborne: free. I will throw this out there too, before we 880 00:51:57,420 --> 00:52:00,870 get into the agenda. But, you know, we talked a little bit 881 00:52:00,900 --> 00:52:03,600 early on before we hit the record button about about 882 00:52:03,600 --> 00:52:06,000 DevOps. And if it's even something that would would 883 00:52:06,000 --> 00:52:11,400 resonate with the users here. You know, it's DevOps in 884 00:52:11,400 --> 00:52:16,590 practice ends up being kind of a collection of practices and 885 00:52:16,590 --> 00:52:22,440 methodologies around product flow. You know, things are out 886 00:52:22,440 --> 00:52:26,130 there's a lot of people and process side to it. And so, you 887 00:52:26,130 --> 00:52:28,980 know, with, would you be interested in the audience of 888 00:52:29,040 --> 00:52:32,700 listening to practices that work, don't work around, you 889 00:52:32,700 --> 00:52:35,730 know, moving things like institutionalized codified 890 00:52:35,730 --> 00:52:40,680 processes into automation, or how to what works and doesn't 891 00:52:40,680 --> 00:52:44,640 work with things like cross functional teams, ways that you 892 00:52:44,640 --> 00:52:48,210 can avoid common pitfalls that we see in organizations, those 893 00:52:48,210 --> 00:52:51,300 types of things? Or are you mostly focused, we'll put a 894 00:52:51,300 --> 00:52:54,840 Twitter poll out, you can email us as well, or DMS, whatever. 895 00:52:55,650 --> 00:52:57,930 But or you just focused on the tooling. And you want us to just 896 00:52:57,930 --> 00:53:02,070 kind of stick with the more I guess, tangible things in 897 00:53:02,070 --> 00:53:05,010 technology, you know, that's a, we're gonna leave that up to 898 00:53:05,010 --> 00:53:08,070 you. If there's, if there's interest, we'll do it. If not, 899 00:53:08,310 --> 00:53:11,580 we'll just we'll skip over it. But I think the next episode, 900 00:53:11,580 --> 00:53:14,280 we're gonna go into ci, and some of the things from a depth 901 00:53:14,280 --> 00:53:18,180 perspective, like code quality and testing and registries and 902 00:53:18,180 --> 00:53:21,240 some of the open source projects that are getting pretty big 903 00:53:21,240 --> 00:53:26,310 adoption there as well, especially from the from the dev 904 00:53:26,730 --> 00:53:30,180 aspect to it, then we'll move into continuous deployment from 905 00:53:30,180 --> 00:53:33,210 the ops perspective, we'll probably touch on some of the 906 00:53:33,420 --> 00:53:36,660 paradigms that we talked about before, like get ops and open 907 00:53:36,660 --> 00:53:39,660 policy agent, but more high level and where that fits in. 908 00:53:40,140 --> 00:53:43,200 And then of course, you can't, can't talk about ci CD right now 909 00:53:43,200 --> 00:53:45,930 without talking about dev sec ops and adding the security 910 00:53:46,260 --> 00:53:52,440 aspect to it there. What we're seeing in terms of adoption, and 911 00:53:53,490 --> 00:53:57,660 what kind of practices and policies are easy to implement 912 00:53:57,660 --> 00:54:00,060 in ci, CD, which ones are harder and where some of the pitfalls 913 00:54:00,060 --> 00:54:03,960 are. And then we're going to build our own pipeline 914 00:54:03,960 --> 00:54:06,930 throughout this. And we're each going to pick your own tools and 915 00:54:06,930 --> 00:54:10,200 talk about what we liked and didn't like, and, and we'll, 916 00:54:10,230 --> 00:54:15,390 we'll show that off. And I dare say, Well, I guess it's we're 917 00:54:15,390 --> 00:54:18,660 open source, we get to put it out there. But we'll have to 918 00:54:18,660 --> 00:54:21,360 document our code this time, I guess, well, we'll do it just 919 00:54:21,540 --> 00:54:24,630 just for this one time, we'll document that code. And we'll 920 00:54:24,630 --> 00:54:27,720 put that up there. That's our plan. If you'd like to see 921 00:54:27,720 --> 00:54:31,020 something different. You know, we're feedbacks always important 922 00:54:31,020 --> 00:54:34,860 to us. So feel free to reach out. Thanks a lot, everybody. 923 00:54:35,040 --> 00:54:35,790 Appreciate it. And 924 00:54:35,790 --> 00:54:36,750 Jamie Duncan: thanks for joining us again. 925 00:54:37,380 --> 00:54:39,630 John Osborne: Thanks, everyone. Stay safe, Happy Thanksgiving.