Recruiting DevOps engineers is hard. It is hard to find them, expensive to employ them, and hard to keep them.
This is a serious issue. In fact, this article is triggered by our recent conversation with the CIO of an American startup whose operations drastically slowed down following the departure of one DevOps engineer. Many companies share the same challenge:
Thankfully, there is a lot you can do — besides breaking your budget. Automate to the maximum, upskill and reskill your existing team, eliminate duplicate work, and stop reinventing the wheel. In other words, get a DevOps platform. More precisely, get a DevOps value stream delivery platform (DevOps VSDP). That is the terminology coined by Gartner, who recognized CodeNOW as a Representative Vendor in his recent market guide (get your complimentary copy!).
Did that sound mysterious to you? Actually, it is simple. A DevOps VSDP empowers your existing team to do more, so you need to recruit less. Pick a DevOps platform as a service — don’t build it yourself, that’s many expensive engineers that you won’t have to pay. Read on.
Several interesting factors make it particularly challenging for companies to beef up their operations team, at the same time that they are facing increased market pressure to improve their IT operations.
First, DevOps is no longer an experiment. Everybody is doing it. That means your competitors too. The 2021 Upskilling Enterprise DevOps Skills Report indicates the popularity of DevOps, with two-thirds of respondents following DevOps methodologies:
DevOps has entered the mainstream
With the broader adoption of DevOps, the demand for DevOps engineers has skyrocketed. This is further complicated by the fact that DevOps engineers are, by the nature of DevOps practices, engineers with strong multidisciplinary skills. DevOps engineers’ skills cover development and operations but may extend to include security, testing, and communication skills. As a result, DevOps engineers are experienced engineers that cannot be mass-produced by the education system. The pool for IT departments to choose from is thus smaller. The IT recruitment agency Ventury explained:
At the time of writing, a quick search on LinkedIn reveals 617 qualified candidates. Of these, only 98 are ‘open to new opportunities’. On the other hand, searching for “DevOps Engineer” on Indeed.com returns 250 vacancies. The problem is obvious – this job market has a candidate : vacancy ratio of roughly 2.5:1. I don’t know of any other areas of tech with a ratio this low. DevOps Engineers have the luxury of being able to pick from a wide range of exciting job prospects.
The relative scarcity of DevOps engineers results in high salaries. Companies that cannot afford the related labor expenses can find themselves priced out of the market. This is a complaint that we have heard often this year from startups, independently of the stage of their fundraising. They can’t pay for as large a DevOps team as they would want to if they are to scale and keep IT costs under control.
An aggravating factor is that DevOps engineers may prove hard to keep around. As Ventury mentioned:
The reality is that DevOps Engineers don’t stay in one place for long. Over 70% of the DevOps Engineers on LinkedIn have been in their current role for less than two years, which means companies have to hire more regularly.
In short, it will not get any better anytime soon. To ease your recruiting needs, you need to simultaneously keep the team you have around longer and maximize their productivity. So how do you do that?
Net result: empowered Devs in build-run teams, Ops/DevOps engineers focused on the interesting and gratifying work, and as importantly, less spend. Let’s start with the first point.
The first lever to reduce your recruiting needs is to increase the output of your existing teams. The DevOps approach revolves around increasing output by improving the flow efficiency of the delivery pipeline. You want to ship new digital applications and services faster than your competitors; improve on existing products quickly thanks to a short feedback cycle; and do so cost-efficiently, without increasing the risks of your operations.
DevOps predicates the importance of organizational structure in achieving flow efficiency. As the name indicates, DevOps recommends breaking down silos between Dev and Ops teams by establishing build-run teams. The practice dramatically reduces communication, coordination, and collaboration time. Non-value-adding tasks thus become more apparent and can be eliminated. But that is only the first kata.
After a decade of applying DevOps, best practices for team organization have emerged and are actively implemented by the best performers. The influential book Team Topologies: Organizing Business and Technology Teams for Fast Flow emphasized the importance of teams as the unit of delivery and identified four types of teams (emphasis is ours):
We highlighted that efficient software delivery requires a team-first approach that relies on long-lived autonomous teams achieving fast flow. […] Team variations [are reduced] to four fundamental team topologies: stream-aligned team, enabling team, complicated-subsystem team, and platform team. When used with care, these are the only four team topologies needed to build and run modern software systems.
Stream-aligned teams are aligned to a single, valuable stream of work; this might be a single product or service, a specific solution or solution subset, a single set of features, or a single user journey. Stream-aligned teams are autonomous build-run teams that independently deliver value to customers with minimal coordination with other teams. They cover the full delivery cycle, with cross-functional skills spanning development, QA, security, operations; they know their customers, adjusting to their feedback in quick development cycles; and they usually follow the Agile principles.
Stream-aligned teams should be the primary type of teams in an organization, while the other types of teams aim at supporting the autonomy, velocity, and focus on customer value of the stream-aligned teams.
A complicated-subsystem team features specialists in a particular system or area of knowledge that is too complicated or too specific for stream-aligned teams to develop in the context of their daily activities. Similarly, the enabling team consists of specialists in a given technical or product domain who make informed suggestions about tooling, practices, frameworks, and ecosystem choices that affect the application stack.
Platform teams provide an optimized product: a DevOps Value Stream Delivery Platform (DevOps VSDP) in the words of Gartner. The productized platform is sometimes also referred to as an Internal Developer Platform. The capabilities of the VSDP are used by numerous stream-aligned teams with minimal overhead. Platform teams especially focus on the usability of the platform via self-service capabilities and supporting documentation.
For instance, stream-aligned teams may use a platform to provision their own environments and resources in the cloud with full autonomy, while the platform team has the responsibility of designing the cloud infrastructure process. The approach cuts out any waiting time, friction, or hard dependency on any separately managed infrastructure team.
4 types of teams for optimized value delivery
With stream-aligned teams, supported by enabling, complicated-subsystem, and platform teams, digital leaders deliver features multiple times every day to their users, recover from disaster quickly, and have a strikingly low rate of incidents. Flow efficiency, DevOps practices, and cloud-native technologies enable all that.
In summary, we described how to adjust the structure of your development operations to improve their productivity. The stream-aligned teams generate value for your customer. They generally focus on your core business. You want to keep that as close to you as possible. As we are now going to see, platforms are the next opportunity to optimize your engineering resources, and most of the time, you want to get that as a service, as opposed to using ad-hoc, quickly assembled in-house solutions. The gains when it comes to recruiting and cost savings are simply too large to ignore.
This brings us to our next productivity lever: end-to-end automation, enabled by standardization. The Team Topologiesbook details:
This approach [e.g., using a platform] has been successfully adopted by many internet-era organizations. The platform team’s knowledge is best made available via self-service capabilities via a web portal and/or programmable API (as opposed to lengthy instruction manuals) that the stream-aligned teams can easily consume. “Ease of use” is fundamental for platform adoption. […] A good platform provides standards, templates, APIs, and well-proven best practices for Dev teams to use to innovate rapidly and effectively. A good platform should make it easy for Dev teams to do the right things in the right way for the organization.
What are the well-proven best practices? In the delivery section of the software lifecycle: Continuous Everything (e.g., continuous integration, continuous deployment, continuous monitoring), Everything as Code (e.g., Infrastructure as Code, Policy as Code, Security as Code, Workflow as Code), configuration management, release orchestration, and more. In the implementation section of the software lifecycle: microservices, containers, service meshes, online IDEs, source control management, event-driven architectures, APIs, and more. In the monitoring section: production metrics measurement, logging, distributed tracing.
DevOps VSDPs implement cloud native’s best practices
When you can get all of that from your DevOps VSDP, your stream-aligned teams can leisurely focus on features instead of repetitive, automatable delivery tasks. You may find out that you no longer need that extra DevOps engineer because you already have 80% of the automation that you need so your existing teams can cover the rest. As many DevOps practitioners, you will find out that automation gives you delivery speed, and that speed goes hand in hand with stability. Paradoxically, as the DORA research team has shown, delivering features faster correlates with fewer defects and rework. And that also means that you are less likely to need extra engineers to put out fires.
It works. It works so well that Gartner forecasted that DevOps VSDPs will be mainstream in 2 to 5 years. Gitlab’s 2021 survey concluded that DevOps teams were “adding the big guns: “SCM [Source Control Management], CI/CD [Continuous Integration/Continuous Delivery], a DevOps platform, and automated testing” (emphasis is ours).
Amazon, Google, IBM, Microsoft, Oracle, GitHub have their own custom in-house DevOps VSDP. Presumably, they feel less the scarcity of talented DevOps engineers. Moreover, they spent years experimenting and honing their platform. In fact, a lot of the best practices that have been identified were popularized by them. You may wonder, if it is already difficult to find regular DevOps engineers for your operations, where are you going to find the highly skilled ones you need to develop a platform? Our answer: don’t! Get your value stream delivery platform as a service.
With a DevOps platform as a service, you do not need the platform teams that you cannot afford. If you do have platform teams, you can still use your existing team to focus on areas that are not covered by the DevOps platform you chose. You can use that team to customize the platform to your own specific business context. Are you in banking? You can add a compliance module to the platform if it does not already have one. The point is that wherever you are going, you will go there cheaper and faster if you build on the work of others.
There are valid reasons to deploy your application on-prem. rather than to the cloud. There are valid reasons to manage your own infrastructure. First of all, you may not have the choice (e.g., for regulatory or privacy reasons). Second, if you do have a highly skilled and plentiful team of specialists to operate your infrastructure, it may be cheaper in the long run than running on Amazon.
In the same way, there are valid reasons to build an in-house platform. As we mentioned, Amazon did. However, for the vast majority of companies, which do not have easy access to scarce engineering resources, nor the time, nor the will to be distracted from their core business, outsourcing the implementation of delivery best practices to DevOps platform specialists like CodeNOW is a smart move. No CAPEX, low OPEX, no recruiting DevOps engineers, and you start on Day 1. This is what Adam Horzenberger, Group CTO at the logistics company Packeta recognized (emphasis is ours):
„We’re facing a scaling challenge. The business grows with incredible speed and our IT systems must keep pace. We’ve started the transformation to a microservice and product-based organization. With CodeNOW, we were able to launch this mission instantly. It’s a solid cloud-native backbone.”
Value stream delivery platforms also decrease the total cost of operations — not just the cost of your human resources. How does this work? If you have 800 devs organized in 120 stream-aligned teams without any infrastructure support, those teams are bound to duplicate engineering efforts. That’s 120 times solving configuration management, high availability, and disaster recovery — often with a wide set of high-priced, proprietary vendor solutions that are incompatible with each other, and create technical debt.
So small companies get access to best practices that they would otherwise simply be unable to implement. Larger companies use DevOps VSDPs to scale their operations while keeping costs from spiraling. There is more. DevOps VSDPs may very well help you reduce developer churn — our next lever to reduce your recruitment needs.
Motivating your engineers to stay longer in the company has many advantages. As mentioned, that decreases your pressure to recruit new skilled staff. As importantly, that is domain, technical, and company knowledge that stays longer in the team.
So how do you keep developers or DevOps engineers? Competing for talent based on salaries only is an expensive endeavor. You need to create a development environment where developers are happy and fulfilled. That is harder to copy and replicate. Three initiatives you can take — which additionally fit well with a DevOps philosophy that emphasizes constant improvement — are to train, grow, and rotate. Choose a DevOps VSDP that will be instrumental in all of these initiatives.
CodeNOW integrates 40 open source technologies into its platform: Kubernetes, Helm, Ansible, ArgoCD, and many more. CodeNOW platform users do not however need to master all those technologies. At the basic level, they do not need to know about them, and they can start coding and shipping applications to production on Day 1. At the advanced level, they may sporadically need to adjust a few variables in a tiny number of configuration files. So what is the training that is needed?
Cloud-native technologies are new to many companies. Deployment and operations concepts are new to many developers. CodeNOW, like any good DevOps VSDP, empowers developers to take ownership of their software in build-run teams. They design it, they code it, they ship it, they run it, they fix it. CodeNOW provides a common platform for all developers to experiment and learn how to configure and operate cloud-native applications in a sandboxed environment in a real cloud. Actual learning of operating cloud applications does not happen on your local machine.
CodeNOW is cloud-agnostic: you can run your application on infrastructure from AWS, GCP, Azure, or Digital Ocean. CodeNOW is multi-cloud: you can run your application on several of those. So developers can be trained to cloud-native patterns without being caught in the maze of vendor-proprietary offerings. Those cloud-native patterns will be useful to them everywhere. In fact, they will be able to acquire any vendor-specific skill faster once they have the concepts in place. So it is not just onboarding that CodeNOW makes easy, it is also upskilling and reskilling.
One of our customers with hundreds of developers on its payroll reported an additional major benefit of CodeNOW. This large company had over a hundred teams, each stream-aligned team averaging 9 developers. Each team focused on a relatively narrow piece of the value creation and achieved mastery of the subdomain related to their stream. After a given amount of time, the company noticed that developer motivation was decreasing.
In response, the company rotated developers between teams. This was made easy because all teams were using the same unifying platform. Developers could be productive in any team on Day 1. This achieved two unintended outputs: cross-pollination of domain knowledge across the entire organization, and happier developers.
CodeNOW decreases your recruitment needs
David Svejda, the Head of New Digital Bank at Komerční Banka (KB), described the value of adopting CodeNOW as a unified value stream delivery platform in the following terms (emphasis is ours):
The standard development platform helped KB to start cloud-native development based on robust unified grounds at scale… It is our platform of choice for long-term, business-driven, technology-enabled transformation.
You can’t find DevOps engineers? Get a DevOps platform, as a service. You will be able to increase the productivity of development teams, you will keep your current developers and DevOps engineers longer. You will find out that maybe you don’t need those extra engineers you cannot find in the market. You will make your current developers happier and grow them into DevOps engineers for a fraction of the recruitment cost of new members. Your existing DevOps engineers will focus on the high-value, business-specific items, which is what you (and they) want.
Are you still skeptical? Talk to us! We can show you our platform. You can ask any questions on our Discord channel. Let’s start a conversation!