Reflecting on the DevOps Role: What It Means for You, Me, and “Steve”

I hope you will find this post helpful, whether you’re a recruiter, or if you’re trying to ascertain your own position within the DevOps definition, or if you need to have that ‘DevOps guy’ in the team, then keep reading on;

Let’s see what the fuss is all about. Thoughts on experiences are my own.

It comes as no surprise that Hiring for DevOps is hard.

In the ever-evolving world of technology, — it has a profound meaning for some, while it means nothing for others.

It is indeed probably a bit of Both.

There are many particular Physical and Virtual, hybrid infrastructure architectures, OS-family-based infrastructures, then there is a myriad of build release pipeline implementation options. Then there is a number of configuration automation frameworks, and now there is a new kid on the block. The Cloud, with own set of product offerings particulars and caveats, with underlying technology stacks, namely Docker, and more recently Kubernetes product offerings. This makes VMs look like the DVDs look too young generation of 2019, — “It works, it’s great, but why do I need these accessories, when I can just stream”.

When it comes to looking for “DevOps guy” it can basically mean… “anyone who has worked with, on, any-of-the-above particular technologies”, — the IT version of “Jack-of-all-Trades”.

So who are we after?

The Definition Wiki says (en.wikipedia.org/wiki/DevOps): —

From an academic perspective, Len Bass, Ingo Weber, and Liming Zhu — three computer science researchers from the CSIRO and the Software Engineering Institute — suggested defining DevOps as “ a set of practices intended to reduce the time between committing a change to a system and the change being placed into normal production, while ensuring high quality”

A formidable definition, but from a Role Title perspective, that would inherently encapsulate a number of computer science derived professions, to achieve the overall objective; an infrastructure engineer, a system engineer, business analyst, product developer, a tester, and probably yet a few more roles I will omit here. Again, who are we, you looking for?

The Technologist The root ‘cause’ of this generic term which has proliferated the likes of LinkedIn for all the best intentions is to mobilise the agile attitudes towards problem-solving and technology setbacks, enabling rapid fail-fast processes, which are to be test-tried to tune to the desired technological productivity metric; whether an efficient QA process, fast Deployment, thus shorter turnaround of Code-to-Customer, all to optimize the systems and for the technology-driven business to remain competitive.

In simpler words, — There rarely is the mere Linux Engineer, or the mere System Admin, who are retired to the corners of the office watching the blinking lights on that monolith system, built some long-long time ago, as their full-time job. Even if so, the very same engineers will likely be in charge of the networking (Cisco CCNA/CCNP), the data storage experts (often with bespoke training on troubleshooting and maintenance — hello HP/NetApp ^.- ) and dabble in ad-hoc shell, Python or other configuration framework scripting, where the need may be.

Alternatively, there are ample software developers who, thanks to the evolution of tool sets, are able to develop and deploy the software, in The Cloud. Armed with the foundation knowledge of the relevant Cloud Provider, the very same engineer can create CI/CD release processes, from the repository, through to production releases. No need for these very same expert storage training knowledge or networking qualifications!

This DevOps Role Title is very much ambiguous, and it all really does depend on you realising that very fact.

• If you’re a recruiting for 'DevOps' — this is to know that not all IT, same as DevOps is the same. And even when it is about-the-same, it may be quite some length apart. Explore candidate experiences in-depth, do query candidates for those relevant examples of work.
• If you find yourself as 'DevOps' or wish to get there — own up your skills and evolve with technology — Moore’s law considered — stay on trend. Easier said than done, especially if your organisation may be operating within a strict technology stack and want you to do exactly what you’re hired for — support and maintain that very stack, at all costs. Make use of the training budget to keep up-to-date, if possible. Mind you, a lot of cloud training portals online are quite accessible, and certification costs themselves are quite accessible nowadays.

My Journey: A Tale of the two DevOps Roles — Self-Hosted vs The Cloud This process is what many classic Linux Engineers/Admins would find themselves to be going through. Adapting to changing the set of tech requirements, In full, — the true meaning of the DevOps associated Agile term.

My own career’s earlier employment was, for a long while, within a large organization. Thanks to ample technical project opportunities, this all resulted in plenty of exposure to a variety of technical challenges and PoCs, which were all opportunities for self-training. That, and being very fortunate to get onto hands-on professional training programmes. The key for me was to deal with those head-on, train up, solve challenges, train some more.

Rinse repeat.

This resulted in my ticking off my First Linux Foundation courses RedHat 6 RHCSA, RedHat 6 RHCE, and the same with RedHat 7 release, then VMware 6 (Cluster Provisioning and configuration) Professional Training, Splunk (Big Data Search) Administrator training and certification

You can guess right, — not all of these are particularly helpful in the DevOps role per se, as far as a generic term is concerned in the all-new, modern, evolved technology stack. But some are useful. VMware in my humble opinion was great, but one day was simply superseded by The Cloud in general. While this was a pivotal self-driven engineering project, the arrival of AWS on the scene the year prior forced me to evaluate the impact of such innovation and what it meant going forward, in my immediate role, and my profession.

Following visits to AWS Summit and Google Cloud Next, I decided to take on the AWS training course and get my first cloud certification. It was time to take the plunge.

Once upon day, - everything changed.

For me, there was hard to resonate with mere “legacy” Data Storage troubleshooting, Networking switch replacements, a multitude of associated hardware inventory-management. Specifically, the hardware-lifecycle management and budgeting, and all-the-while make do ansible-application-configuration, for only some of that infrastructure, — Virtual Machines. The proverbial “Cat was out of the bag”.

It was time to leave and seek a new challenge, — Engineering on, or Migrating, - Into The Cloud.

• I left as a Senior DevOps Engineer, of the old.
• I went for a new role — a Senior DevOps Engineer, yet of all the new to come.

So, Hello Loveholidays.com ^.-, — most tech-forward start-up I know, natively embracing the technology. Pushing the cloud limits, a day at a time, and literally.

“Same words — all new meanings”

In my role at Loveholidays as a Senior DevOps Engineer, a AWS conferences deep dive followed, and SysOps Professional training with certification achieved. And then, yet more Google Cloud Conference deep dives, with Professional Training and Certification followed-on as well, for good measure. After this much cloud platform evaluation process, we eventually opted for Google Cloud. To us, here, now, this all falls within the remit of all the same DevOps. It is Our Company DevOps culture, perspective and thus skill requirement.

DevOps Recruitment, — It’s for the Business “We’re looking to fill the DevOps Role…” It falls down to the Hiring business organisation to help define the role they are trying to fill. Ask for skills needed, basic requirements and advantageous skill sets.

There is no ‘One DevOps Engineer fits all’. It’s a mere catchy generic word which helps search. A given, but rather the devil is in the detail, as we all realise.

Rather than for the general tagline search or advert, — if want a developer-first — ask for it from the outset. It is easy to get lost in this generic expectation and assumption that both sides of the table know that ‘DevOps’ implies, but this can also be time wasting for both sides, resulting in incompatible roles offered/taken.

DevOps Engineer can be any of the following;

• Linux Engineer
• The developer who works on their Mac, and is ‘good at’ Jenkins
• Network Engineer who works on their Ubuntu Laptop
• System Admin who manages [any] CI/CD Pipeline, general IT management
• Cloud Certified professional, who deploys their code using cloud-native product offering or via bespoke Docker.
• Any Permutation of the above.

All of these carry a degree of expertise in their respective immediate remit. Not all can even be tested the same on the interview assessment day without prior proper preparation on the core basics.

Assessing the Candidates With the definitions and ambiguity of the Role Title: What Business requires, versus what DevOps Person Is, out of the way, Let’s talk candidate assessment.

There may be that ‘First Time’ for seeking that candidate for your team, or It’s an expansion of the existing team, a process, which I hope is tuned well to find the right candidate. Either way, the below is my list of assessment guide, which has worked really well, with new-added tips;

• Don’t ambush the contender — again, given the role ambiguity, ensure the candidate knows what particular set of skills you’re looking to fill the position, along with assessment stages — so the candidate can prepare accordingly. Just because you have worked on some ad-hoc Python development project some few years back, and have not touched it since does not mean you’re still that Python Developer today.
• Offer the relevant Homework Test exercise — this can typically allow you to see how apt the candidate completing the set of requirements along with their technical capabilities. It would be most relevant to businesses own day-to-day use case, to be delivering a working solution.
• Technical Discussion should be relevant to Business actual technical requirements the role is trying to fill. I don’t see the value for the University Challenge question type, asking for definitions. Some tools/protocols/abstracts have become so ingrained in our daily ‘DevOps’ lives that put on the spot, answers to which do not prove or disprove candidate’s actual ability to effectively perform their job. If it is a release pipeline modelling, then ask those very relevant questions. Resort to white-board where necessary. The candidate does not need to know where/how to view manuals for a particular binary or their extended use-cases, — when such manuals are better googled for examples, tried to be fine-tuned. Do You know the full set of netcat command options, and when to use those? Or how would you do a ping check via a particular interface? When put on the spot, with the minute count to the interview, these are irrelevant questions, and time is better spent elsewhere, talking best practices, approaches to problem-solving, and so on.
• Prepare a technical challenge — If Home Test is not possible, there is always an excellent consistent standard for all your candidates to meet, the same test you put through all your candidates, before deciding on the team fit. This is something which would not be longer than an hour. You should always do the test yourself, first, before passing it to the potential candidate, even if to flag any issues with the assessment system(s)

The Candidate If you are ‘DevOps’, that’s great! — now try to ensure you’re most up-to-date on all the latest technology stack. Even if your immediate role/business does not get the exposure to it.

It’s wise to be in the know. Just like a good security patch, — it’s in your interest to be up-to-date, — y’now, for security.

If You want to get into ‘DevOps’ — Good places to start are the initial conferences on the topic — #AWSsummit or #GoogleNext (you can watch conference sessions online on youtube) to get the basics, then follow-through on BreakOut Sessions on relevant material. Wrapping up with the relevant training material and a certification to boot.

My DevOps Interview Guide I personally view, The DevOps role, as a good Linux system administrator as a base foundation first, with some scripting/automation experience. A more advanced System Engineering type for more senior positions, with an effective configuration management technical requirements via tools of choice, following best practices, while doing so. In the new world, you would also need to be up-to-speed on Docker containers, and best practices there as well.

Bash shell scripting is common, in both Linux and micro-services world alike, so be prepared to answer some question in that regard also, typically for a more senior-tailored role.

Prepare for Interview with a good recollection of projects, which solely yourself have undergone through the painstaking process of delivering a solution, end-to-end. The challenges you have had along the way and the technical solutions you’ve come up with. It’s helpful, to be frank. Interviewers will typically be another technical savvy person who has also delved into the pain and came through it.

Be Prepared to whiteboard some basic concepts or at least to talk through those at length, — CI/CD — what is it, how it works, how would You implement it, what options. As an engineer-first, how would you do it? Why?

Getting Ready to be ‘DevOps’ Generalist Most of the training on is easily accessible online these days, so there is no reason not to be aware of what’s out there, here are some of the links to my personal favourite go-to training material essentials;

• Cloud Training and Certifications thanks to www.linkedin.com/redir/general-malware-page?url=https%3A%2F%2Facloud%2eguru%2F — one stop shop — a great place to get started towards all cloud certifications.
• Docker, Kubernetes Know-How — I must thank Nigel Poulton for great deep-dives on the topic with easily digestible material via Pluralsight www.pluralsight.com/authors/nigel-poulton
• Linux Basics — get industry standard Linux training — RedHat RHCSA prep course via Udemy — one of the recently published courses www.udemy.com/rhcsa-training-class-rhel7
• Automating Configuration — with Ansible, my bias legacy preference of course -www.ansible.com/resources/webinars-training (There are Puppet/Chef/Terraform for cloud provisioning) It’s an essential component to provisioning consistent systems.

Once training and certifications are ticket boxed, congrats — you’re in DevOps business!

Now, apply for the relevant roles and good luck.

Why to stop learning there, there is a Technological Zen — infrastructure-as-a-code, — based on Docker + Kubernetes technologies, I strongly recommend you to be looking into.

On That note, I hope you enjoyed the read.

Best of luck on both sides of the table, and may the best ‘DevOps’ win.

Like, Comment, and — sharing is caring.