DevOps Digest #2: Can outsourcing DevOps services work?

Welcome to DevOps Digest, brought to you by cloud consultancy, Vivanti. A news and analysis podcast, we cover the trends, thought leadership and announcements happening in today’s DevOps space. Topics span from Continuous Integration and Continuous Delivery (CI/CD), Microservices and Infrastructure as Code; to Monitoring and Logging, App Replatforming and Pipeline Automation, Collaboration and more.

Episode two focuses on the current state of the tech hiring market in the US, particularly the dearth of available DevOps talent. We discuss surveys and news related to five areas:

  1. The demand for DevOps talent and the drivers of that demand
  2. How can organizations address today’s gap between supply and demand? Is outsourcing DevOps services the answer
  3. How to attract DevOps talent over the long-term
  4. What constitutes good DevOps at an organizational level
  5. The skills and attributes most important to today’s DevOps professionals

DevOps Digest, the second coming

Lachlan James

Alright, hello and welcome to DevOps Digest, a news and analysis podcast about everything DevOps brought to you by cloud consultancy, Vivanti. I’m your host, Lachlan James. Vivanti Principal Consultant and DevOps guru, James Hunt, has decided to join me for round two. And, he’s brought a friend with him: Enterprise tech stalwart, Mike Walker. We also have Raj [Iyer], our new Principal Consultant for Data. Good to see you gents. James: Are you as surprised to be back for Episode Two as I am?

James Hunt

I was surprised YouTube didn’t pull the whole thing.

Lachlan James

Yeah, look, I was thinking about [that possibility too]. But hey, here we are. We’ll see if we can do better this time around [laughing]. Mike: Thanks for jumping into the hornet’s nest. Now, James thought we were getting stale after one showing. So my question for you is, how do you plan to liven things up for Episode Two?

Mike Walker

I think we’ll keep it a little shorter than Episode One. I think that’s all we can do…

Lachlan James

That’s a very serious answer there, Mike. I was hoping for a comedy routine. But hey, here we are. He’s playing this with an extremely straight bat. But he’s Englishman, so that makes sense [because he hails from England, the home of cricket]. It won’t make sense to any of the listeners, but it makes sense for us.

Demand for DevOps professionals is soaring

Lachlan James

Alright, I’m gonna dive into the first sort of section we’ll talk about today. We’re going to zero-in-on the demand for DevOps talent, and within the developer market more generally, and how those things are relating. So as we all know, the tech sector has been on fire. And whilst we’re likely seeing the beginning of what’s a cresting wave, with some venture capitalists taking their money and heading for the hills, good help truly is hard to find right now. And to give a little bit of context, before I sort of throw over to you James: Business Insider wrote a recent article [Big Tech Salaries Revealed], which revealed that the US tech sector experienced job growth in 10 out of 12 months, last financial year, adding 80,000 workers. And, this is at a time, you know, when the economy was shedding jobs elsewhere — so it’s a pretty outstanding result there.

Big Tech salaries revealed (Business Insider)

More recently, the Computing Technology Industry Association published new survey data last week, and basically they were reporting that job postings remain at record levels, and technology companies added workers for the 18th straight month. So narrowing this down, CodinGame’s 2022 Tech Hiring Survey put DevOps roles as the third most sought after role in software development. And that’s pretty massive, considering that only 2.3% of survey respondents said that they currently worked in DevOps roles, yet, you know, the hiring data in that same survey suggests it’s now the third hardest position to hire for within software development. So, James, what’s happening here? I know developers have been getting snapped-up like yesterday’s leftover pizza. But, why does it seem that demand is way outstripping supply, particularly for DevOps?

CodinGame’s 2022 Tech Hiring Survey (CodinGame) DevOps third in 2022 Tech Hiring Survey Top 5 positions recruiters will struggle to hire for in 2022
James Hunt

Well, so for starters, the CodinGame [survey], they actually reported on two different axes for the demand slash prioritization, right? DevOps came in second, not third. Web Dev was the most demanded by recruiters, followed by DevOps, followed by AI and ML, which is really surprising. The developers themselves prioritize their own desires to work in Web Dev, followed by Game Dev followed by DevOps. So they’re close. But DevOps is actually even more important to the overall digital economy than the article originally leads you to believe.

Top in-demand skills for 2022

The reason I think is because more and more companies are going digital. So while we’re dropping back office personnel, we’re dropping physical jobs, we’re moving more and more things into the cloud. More and more things into computers. And it’s hard to believe we’re still doing that and 2022. But, that’s what I think is going on.

Lachlan James

No, that makes some sense. Mike: I’d like to pull you in here too. And just [provide a few pieces of contextual data]. On June 3, Global Industry Analysts promoted their latest DevOps Global Market Trajectory & Analytics report. And it forecasts the DevOps market to basically double from around almost 9 billion this year to approaching 17 billion by 2026.

A $16.7 Billion Global Opportunity for DevOps by 2026

And we’re seeing this play-out locally and specifically too. During the week, for example, we had Silicon Angel [DevOps darling GitLab’s stock soars as it easily beats earnings targets], capital.com [Gitlab (GTLB) earnings propel the DevOps stock’s Tuesday gains], and The Register [GitLab spots huge opportunity for DevOps platform as revenue soars], all drooling over the better-than-expected first quarter earnings data from GitLab. And so in fact, you know, revenue jumped 75% compared to the equivalent quarter last year.

First quarter earnings data from GitLab

So Mike: How do these figures play into the rapid rise of DevOps? I mean, GitLab’s CEO suggests it’s basically because they have no real competition. Given the global figures we’ve just cited too, is that a bit overblown? And is this indicative of something else?

DIY DevOps vs GitLab
Mike Walker

I mean, that’s an amazing sweeping statement, that they have no competition. I think we all know they have at least one really sizable competitor. So I’m not sure how that got past him. But to answer the question, Lachlan, yeah: There’s a clear demand for more rapid delivery of services. And I don’t think there’s an IT department out there that isn’t doing everything it can to increase process automation. So if you combine that with the report that you cited that says, you know, this increased demand for PaaS, growing needs within the enterprise to drive operational agility and customer satisfaction, I think you have your answer right there.

Lachlan James

Yeah, no, I agree. It was just some interesting specific news, in relation to that trend, that came out this week. And, some very interesting comments from [GitLab’s] CEO there. I mean, obviously, they’re trying to position themselves uniquely. And, I understand that, but I thought that was very interesting, in amongst a lot of global data that’s coming out at the same time. But no, I think that’s very fair to say, it’s obviously part of a much bigger, broader trend, as James was saying, as everything’s getting into the cloud. I mean, DevOps [has been around for], what is it, 15 years now? But people are really understanding how important it is now. So I think there’s a mad scramble, going on within the mad scramble, that is tech. So I think that makes a lot of sense.

How can organizations address today’s gap between supply and demand?Is outsourcing DevOps services the answer?

Lachlan James

So this, to me, obviously, takes us to another question about addressing the skills gap. So you know, taking what we’ve discussed so far, as I said, the next obvious question is, well, if there’s a widening gap between available DevOps talent and market demand, what can organizations do to combat that? Because it really is a growing problem. For example, the DevOps Institute released its 2022 Global Upskilling IT Report, which forecast an 85 million person shortfall, and $8.5 trillion worth in unrealized annual revenue, should this trend continue that we are seeing.

2022 Global Upskilling IT Report

And, before I come to you James, I’ll provide a bit more fodder for your cannon, because there’s a bit going on around this space. So TheNewStack recently wrote an article arguing that within this sort of context, you can’t hire your way out of the current cloud technology skill shortage that we have. Because, as they were saying, there’s literally not enough talented professionals in the market right now to do that.

You Can’t Hire Your Way Out of the Cloud Skills Shortage

So TheNewStack piece essentially offers four alternative approaches. They said:

  1. Hire under qualified individuals and train them;
  2. Advertise training opportunities when recruiting;
  3. Take advantage of certification exams; and
  4. Sponsor scholarship programmes.

Now, to me, they sound like reasonable strategies. But, it’s hard to see any of that bearing fruit in the near-term to help solve immediate needs. Alternatively, a recent report on the DevOps Outsourcing Market said this was an increasingly popular option. But, you have to be really careful here. So James, can outsourcing work for DevOps? And under what conditions?

James Hunt

I definitely think outsourcing in the DevOps space is a great short-term strategy. Because to your point, yes, hiring people and training them up to be where you want them is a very valid business tactic. Build the workforce you want, right? Hire them at entry level, give them what they need to grow, and they will surprise you. We believe very strongly in that. But that’s a long-term play, right? That’s quarters and years, not weeks and months. To me, the promise of DevOps and automation is actually an eventual reduction in the amount of work. The reason we digitize business processes is because digital systems are a lot more flexible, easier to extend, expand, scale and change over time. But to get to that point, requires a fair amount of work. And to handle that, we do what we do whenever we have some scarce resource – whether it’s money or time – we take out a loan. And that’s really what the principle of technical debt is in DevOps: Let’s do something that’s not perfect today, so that we can get somewhere that’s good enough now, and we’ll come back and we’ll make it way better later. In order to pay down that debt, it takes some ‘bursty’ work. You have to be able to scale-up a team to go tackle all that debt, tackle all the things that you know are tucked away as ‘fix-mes’, and ‘to-dos’. And we should really come back and see if this is the right architecture. Or, let’s finally figure out how to use this API or that tool.

James Hunt

To do that, I think outsourcing is a perfect way. Because a lot of the same technical debt is everywhere, right? Lots of companies, they’ve cut the same corners, they make the same short-term decisions to get the velocity up. And then, having a consultancy or contractors come in who have seen this elsewhere, might actually make them go faster at resolving that technical debt.

Lachlan James

No, that’s actually a really interesting take. I wasn’t fully expecting you to say that, James. But, that’s what I like about you, you keep things interesting – because somebody has to! On that note of super-interesting, Mike, that’s why you’re here. Do you have a view on this?

Mike Walker

Really only to say that, you know, the real benefit of outsourcing DevOps, if you’re trying to build a practice internally, is to basically bootstrap that team or bootstrap that function inside the company. And, to James’ point, as you’re trying to grow those resources internally and get to that stage where you become self-sufficient… Having a team that you can outsource to, and then work alongside you and your people, is probably a reasonably effective way to do that. Again, reiterating what James said, you are then able to adopt best practices that have been learned by essentially consultants that have done this in many different places in many different industries – hopefully one very similar to your own. [You can then] adopt those [best practices] and bring those in-house. So, better to learn your mistakes through proxy of others, than it is to learn them on-the-go yourself. So yeah, I do believe it’s an absolutely viable thing to do. I think it’s probably one of the most efficient ways to rapidly grow [your] DevOps practice internally.

Lachlan James

Yeah, and I think, as we were saying, for near-term solutions, that makes sense. There are lots of other things you can do in the longer-term. But, for now, we know there’s an issue, and that won’t be solved really quickly, because of this really big gap between supply and demand. So I think that’s an interesting comment.

How can you attract a steady stream of DevOps talent over the longer-term?

Lachlan James

So taking that, I think this leads us to an interesting problem, looking at it slightly differently, right? So from, maybe, a more strategic perspective. We’ve just talked about some options there to augment your in-house team. But, how do you effectively attract a steady stream of DevOps talent over the longer-term, right? Because obviously, there’s what you can do right now. But, you should be looking strategically and thinking about what you do to actually change the status quo and work effectively over the long haul. I’ll get the obvious things out of the way in regards to that sort of longer-term strategy…

Lachlan James

We know we have to pay people well, right? FinExtra journalist, Barry McCall, did an interesting write-up [Salary scales indicate it’s a great time to be in tech].

Salary scales indicate it’s a great time to be in tech

He cited some new data from Robert Half‘s 2022 Technology Salary Guide, which indicated the average DevOps engineer went from about $USD 110,000 in 2021 to $USD 126,000 in 2022, because we know this is a really in-demand sector.

2022 Technology Salary Guide
Lachlan James

But, what else really matters? And I think part of the answer has to lie in equipping people with the ability to deal with the deluge of demands that this overall talent shortage is throwing at them right now. In that same vein, VentureBeat recently wrote about this, [in an article titled ‘How to succeed in digital transformation amid growing talent shortages’], saying that: “Already spread thin, DevOps are having to adapt to the cloud-based needs of today’s businesses.” And, you know, there’s a lot going on, and they go on to say: “Until organizations are able to source adequately skilled professionals, DevOps will need to learn to manage, deploy and ensure interoperability between edge, container, AI, security and other technologies. It’s also important to note that they’ll need to understand when and how to refactor applications and manage multi-cloud technologies, while knowing how to implement and manage applications deployed in the cloud versus on-premises.” And they say, because they’re quite different skills. So, a lot to consider there. And I think, you know, that’s probably pretty on-the-money.

How to succeed in digital transformation amid growing talent shortages

So James, what does this landscape mean for attracting and retaining DevOps talent into the future?

James Hunt

This is a bit of a weird thing to say, but I think people need to be able to fail, right? We’re at an inflection point of technology adoption, where what worked five years ago isn’t working today. And what will be working five years from now is what we’re going towards. But in order to get there, people have to be able to experiment. They have to have room to try things. Sometimes those things won’t necessarily pan out. Sometimes they’ll blow-up in your face. And I think, to pull this off, the business has to take a deeper view of precisely how these solutions come to be. How do we build these cloud-based architectures? How do we architect applications? What goes into making a microservice? Those kinds of things – the more holistic approach, or the more holistic understanding, of what it is that DevOps does all day.

James Hunt

I like to say that no financial analyst builds a spreadsheet by starting at ‘A1’ and just moving along cell, to cell, to cell… plopping in formulas and dropping figures into the cells methodically, right? It’s a thing that’s like a sketch. You start with a basic idea, you go down a couple of paths, sometimes they work, sometimes they don’t. And the same thing happens in DevOps. And I think in order to retain DevOps talent – and really to grow them to where they need to be, to get to the skill-sets we need in today’s workforce, which we want 10 years down the road – we have to be able to foster that play, that experimentation, that failure.

Lachlan James

No, I think that makes a lot of sense. Mike, you’ve led lots of big teams over the years: What’s your view on this?

Mike Walker

Yeah. So, if we’re talking about, how do we attract and retain the talent? I think what it comes down to is, you have to provide an environment that’s appealing. So the work you do has got to be fun. And DevOps is fun, right? You’ve got to give people access to some of the latest tools and give them the freedom to use them, the freedom to fail when learning how to use them. And you need to create an environment that is generally appealing and exciting; pleasant for people to be in. So you remove friction. And by friction, I don’t mean, just general workplace friction between people, I mean, all of the things and all of the administrivia that generally goes with working inside an organization, trying to minimize that as much as you can. Retaining people just really means keeping people interested, keeping people employed, keeping on giving them challenges to solve – if that’s what they’re driven by. And DevOps is fun, because you get to see the results of your work go into effect much more quickly than you would in a sort of traditional development environment. So having that visibility across the whole spectrum, I think is very, very important. And that’s what’s always worked for me anyway.

What constitutes good DevOps at an organizational level?

Lachlan James

Alright. So we’ve talked through a number of different concepts about what’s happening in the market, specifically DevOps supply and demand, and what you should do in the short-term and long-term when it comes to DevOps talent. The next group of articles I’ve recently come across, talk about the impact good DevOps practices can have on organizations. [They] also [introduce the concept of] what constitutes good DevOps at an organizational level. So StrongDM.com published a bit of an SEO stocking filler, titled ‘40+ DevOps Statistics You Should Know in 2022’. But, it did have some interesting stats.

40+ DevOps Statistics You Should Know in 2022

And one of those was a DevOps Trends Survey conducted by Atlassian. And basically, 99 percent of respondents said that DevOps has had a positive impact on the organization. Now, you know, probably a little bit biased, sure, but it’s hard to argue with all-the-same – that sort of response rate. But, when it comes down to the components that make or break that positive impact, there’s seemingly a thousand different views out there. So I’m going to highlight, I guess, a few recent ones, to get your juices flowing. I’ll then throw over to you, James, and hopefully you can work some magic.

Atlassian DevOps Trends Survey 99 percent of respondents said that DevOps has had a positive impact on the organization
Lachlan James

So a couple of takes I saw on this over the last week… EnterpriseTalk.com went high-level, citing ‘setting unattainable objectives’ and resisting automation as show stoppers [in its article, Three Key Reasons Why DevOps Fails]. TechTarget thinks embedding continuous cloud cost optimization into DevOps pipelines is important to underpin sustained DevOps success [outlined in its article, Reduce cloud waste with careful DevOps cost management].

Talking about DevSecOps, DevOpsOnline went deep on clusters and containerization, [in the article, Four must-know principles for DevSecOps], pinpointing four core principles:

  1. Monitoring: So using automated tools to issue manage, and monitor machine identities inside clusters;
  2. They talked about consistency: Clearly defining and communicating straightforward execution processes to ensure more traceability for each new cluster spun-up into production;
  3. They cited identification: So introducing security tooling to automatically scan containers for common vulnerabilities and configurations against security policies;
  4. And lastly, they also pinpointed isolation: So reducing the blast radius of security breaches by isolating applications.
Four must-know principles for DevSecOps

So they went pretty narrow there, whereas TechTarget’s SearchItOperaitons.com offered a roadmap to enterprise-wide DevOps adoption, arguing it all comes down to setting the right success metrics from the start [DevOps goals and objectives for a smooth adoption roadmap].

DevOps goals and objectives for a smooth adoption roadmap

So, look, Raj, what are your thoughts here? What are the factors you’ve come across that have the most impact for success?

Raj Iyer

Yeah, thank you. So I think there is one more angle, which is very important for successful DevOps, which is that it needs to be tied to a higher purpose – to an organizational goal. So it’s one thing to keep it to, let’s say, SRE metrics, and we talk about you know, SLI’s and say, ‘okay, we need to hit those, so we’re going to automate, we’re going to introduce certain things’. But then that needs to be tied to the higher purpose of the organization. I think that is a fairly big motivator, in my opinion, to rally the troops. And then the other aspect of it is giving a level of autonomy – within certain guardrails – to empower the team to innovate. [That is], to experiment in a way to reduce the friction, which Mike was referring to a little while ago. I think, if we can give a little bit of autonomy, give them a little bit of space for experimentation, and tie it to a higher purpose – to the overall success of the organization – that’s the secret to success there.

Lachlan James

Raj, thank you very much for throwing your hat in the ring. That makes sense. James, when you think about success from a DevOps perspective, at an organizational level, what are the things you go, ‘well, you’ve got to have those’?

James Hunt

Before I get into that, the thing I found amusing while I was reading through the research for this… 99% of respondents [to the Atlassian DevOps Trends Survey] said that DevOps has had a positive impact on their organization. I assumed that those who didn’t respond [positively] were trying to salvage a production deployment – from several 100 pipelines, across several dozen unique Jenkins servers – and didn’t have time to fill out the form. Because, DevOps is something that can definitely go very, very far south in terms of mission and where you’re trying to get to. And to Raj’s point, tying things to the business vision, where you’re going, is probably the most important thing. So EnterpriseTalk.com talked about that, right? They said the opposite: Setting unattainable objectives, which is also often something that’s not really in line with what the business wants to do. It’s not unattainable because it’s technically not possible. It’s unattainable because there’s a lot of human process in the way that’s gonna try and stop you. Because that’s not really what we’re trying to build here. That’s not who we are. That’s not what matters to us as a business, right? It might be cool to have deployments go from dev all the way to production in less than five seconds. But, we’re a regulated industry, and we absolutely cannot handle that. So the objectives being in line with the business is the most important part. I did find it amusing that TechTarget went into the cloud cost because, reading between the lines there, what they’re saying is that the only thing that causes DevOps to fail is run away costs.

Lachlan James

Right!? I was like, ‘No, I think there’s a whole lot of other issues before cost becomes a factor’, right? Because if a whole lot of other things are failing and are a disaster… you’re not gonna get to the point where you’re running it for long enough to go, ooh, ongoing costs are high. I agree.

James Hunt

DevOps is going to cost you in human capital: In paying people to do stuff that they shouldn’t do, or doing stuff that they ought to automate way faster, [rather] than paying an extra cloud bill. I’ve worked with customers where we’ve shown them: ‘Here’s your monthly AWS bill. We think if we worked for the next six months, we could cut this down by 10 percent’. And they’re like: ‘We would spend way more on you trying to save 10 percent’, right? Pennywise, pound foolish. So interesting that TechTarget went that way. And, I don’t know what the DevOpsOnline folks are talking about [in regards to DevSecOps], because this is all tech, right? Monitoring, consistency, identification, and isolation is really the tactics of building the process into the computer – which is what I really think DevOps is. I think the thing that makes DevOps really work is belief. Like any human undertaking, and DevOps is 100 percent a human undertaking, it requires everyone involved to be working in the same grain, chasing the same outcome. And the tech that you use is largely irrelevant, except when it stands in your way.

Lachlan James

Yeah, I thought that was super interesting [that] you picked up on that point. Because they [DevOpsOnline] sort of just listed out.. almost just described the [DevOps] process.

James Hunt

This is what it looks like to do DevOps, ergo, this is what it must be to be successful. Like, no, there’s a lot of people doing containers and isolation, and scanning, and then not doing the spirit of DevOps – just like there’s a lot of people who do story points and don’t do Agile. Or, they have a bunch of sticky notes and don’t do Kanban. It’s the human element of that process that matters way more than the tech.

Lachlan James

I almost had a stroke when you talked about Kanban Boards and sticky notes. Because, if you can see the state of my desk, there’s many-a-sticky-note that has no real purpose or home…

James Hunt

You just have them stacked up higher and higher. Right?

Lachlan James

That’s not too far from the truth. Yeah, absolutely. I mean, I don’t think the issue here is people not understanding the components or the processes that make-up an ideal DevOps practice. I don’t think that’s the thing, as you said, it’s human factors. It’s about getting people to work together better for longer, and making that a success. I wholeheartedly agree.

Which skills and attributes are most important for individual success as a DevOps professional in today’s world?

Lachlan James

And so, on that note, I kinda want to move, lastly, back to the individual level. We’ve been talking about the organizational practice of DevOps and success there. So, if you had to pick and choose, what are the most important professional skills that DevOps folks need in order to help them survive today’s world? The same 2022 UpSkilling IT Report, by the DevOps Institute that I talked about, pinpointed five broad brushstroke things when it comes to individual DevOps skills.

The Top Five Global Skill Capabilities

They said:

  1. Process and framework skills (57%),
  2. Human skills (54%),
  3. Technical skills (52%),
  4. Automation skills and automation tool knowledge (49%),
  5. And leadership skills (49%).

Pretty broad but, at the same time, if you go that broad it’s hard to argue with too…

James Hunt

Right? Doing things correctly is a good skill. We should do things correctly [laughing].

Lachlan James

I love lists that are motherhood statements, right? But to be fair, at a high level, reasonable points. And, on the other hand, TechTarget’s Matthew Grasberger focuses on programming languages as the pathway to personal success [in his article, Important DevOps engineer programming languages to learn]. He was emphasizing the ability to:

  • Design CI / CD pipelines with YAML;
  • Learn object-oriented languages, like Python and Ruby
  • And run Infrastructure-as-Code with tools like Terraform.
Important DevOps engineer programming languages to learn

That’s kind of the angle they went with. They went quite deep.

James Hunt

I’d like to raise one point of order: YAMAL is not a programming language. It is a mark-up language. And I don’t understand, because TerraForm isn’t turing complete, in that you really can’t do anything other than deploy. I mean, last time I checked TerraForm, it was fairly declarative, and fairly descriptive and prescriptive. If it’s gained the ability to do arbitrary computation, we really need to stop making these mistakes with our tools [laughing].

James Hunt

So, to my mind, those are the tactical ways that you survive as a practitioner of DevOps. For a very short amount of time, you can get by on just having the ability to write code. You can get by on just having the ability to talk your way out of a story pointing exercise. You can get by, for quite a while, being able to automate the manual tasks that others are doing. But, to me, the long-term success in DevOps is the ability to look at a solution that is suboptimal, which doesn’t do everything you want it to do, and accept that it’s good enough. We understand that it is not the perfect solution. It’s not the most generic; it’s not the most reusable; the most modular. You need to understand precisely where it is sub-optimal, and then – this is the most important part – remember that fact, for at least a couple of years. It’s, it’s kind of like we teach our children to be fiscally responsible. We teach them about money, and how the world works on top of money, and loans and finances and all this. [When it comes to DevOps], we need better technical debt management, if you will. A lot of people will say: ‘Oh, we’re gonna do this, oh, that doesn’t solve our problem. That’s okay. Tech debt, move fast, break things. Great’. But, if you don’t remember what you broke, you can’t ever come back and fix it. And on that note, I’d like to tell you a story, about 17 years of session data.

James Hunt

So I was working on a project where we managed a very old e-commerce platform. This thing has been around since before the.com, bust, right? So late 90s. And we had a situation where the customer called-up and they were having issues getting orders placed. The system was sluggish. And this ended. And we tried a bunch of stuff. And what we eventually found was that they had kept all of the session records in the database from the beginning of time for the system. They had 25 gigs of session data. And it was one of those things that we finally chased it down to a piece of code that had a ‘fix me’ in it. And the ‘fix me’ said: “We’re not currently expiring sessions. But that’s okay, I need to get this into production. I’ll come back next month and fix it.” So, this was in September of 1998. And here we are in 2022. And, if it weren’t for a massive database outage a while back, it would have been far more years of session data. So being able to say: ‘Yeah, I understand that this is not where I really want this code, this project, this architecture, this solution to be’, but then to have the wherewithal to come back and fix that technical debt loan… That to me is the success. The one defining characteristic of a successful DevOps person.

Lachlan James

I remember, in our last podcast, I think the phrase you used was: “There’s nothing so permanent as a temporary solution”. So that’s pretty indicative of that, I think. So, that’s a characteristic. Totally buy that, right? We’re talking about a pretty human pursuit here. Does that idea [you’ve just outlined] boil down to a skill, or a set of skills? Or do you think skills, at an individual level, actually aren’t that relevant and it’s about characteristics and approaches for individual success [as a DevOps practitioner]?

James Hunt

Again, because DevOps is a human thing, not necessarily a computer-robot thing, the skills are relevant insofar as you have to have the skills to do the work, right? To truly excel at the work, you have to go above and beyond. Just like you have a hard time being an actuary, if you can’t do math. But, if you’re an actuary who doesn’t understand what the business that you work for does, you’re not going to be a successful actuary.

Lachlan James

No, that makes plenty of sense. Raj: Where’s it all sit for you? What’s the secret sauce to personal success in DevOps? What are you looking at when you think about individual success and the skills – or, as James said, the characteristics – that are important?

Raj Iyer

Checklists. I think if there’s one word, or one thing I would say, which helps in our DevOps processes it’s checklists; and automation of checklists. And what I mean by that is a more checklist mindset, which is that whatever mistakes I’m seeing, whatever issues I’m seeing, if it should not be repeated next time, I have to add to the checklist or modify the checklist. And to some extent, what James said: It’s just the discipline of coming back and fixing, addressing the debts we are creating. And sometimes issues are procedural, not necessarily technical. So, in my experience, if you want a very succinct, very simple way of improving DevOps, it is getting the checklist mindset. Not necessarily a human checklist, even if it’s system-based checklist verification. If you can keep improving that, if you just focus on that, I think a lot of success can be had.

That’s a wrap

Lachlan James

Okay, that makes sense. I think that’s something pretty practical to end this on. So, we’re probably out of time for this episode of DevOps Digest, The Second Coming, brought to you by Vivanti Consulting. James and Raj, I really want to thank you for your time.

Lachlan James

For those watching, thanks for tuning in. To receive regular DevOps news and analysis, you can subscribe to our YouTube channel or go to vivanti.com/contact-us and sign-up to our mailing list. So everyone, it’s been a pleasure to have your company, and bye for now.