A DevOps person isn't someone who develops, and who does Ops. It's someone who does only Ops, but through Development.
It's not about start ups vs Enterprise, it's about 1 person writing programs or 5 people doing things by hand.
The older, more foundational problems were getting automated back then. Now that they're solved problems, and combined with more and more people running large and/or virtual infrastructure, a new problem domain exists around spinning up machines and deployment.
The current coding investment is infrastructure because it's the current pain point. In a decade (or whenever permanent solutions exists for infrastructure) the current way will be considered "by hand" and operations coding efforts will just move onto whatever problem is only visible now that infrastructure is no longer a time sink.
You can say that some ops is just admins running already existing software and operating everything by hand, but there will be admins doing exactly that in a decade too.
Generally, a sysadmin has slightly different skills from a developer - they might code in a highly imperative style and always keeping the actual machine/system being targeted in mind, but I've never known a half-decent sysadmin who cannot write code.
If that's the term the market wants to use, fine. As far as I'm concerned, a senior sysadmin who can't write in a couple of scripting languages isn't senior.
I know a reasonable amount of sysadmin (all my computers run Linux primarily, I only keep Windows on for checking hardware issues, and a couple of specific apps I need to run once or twice a year).
I wouldn't apply for sysadmin jobs, because I wouldn't feel my knowledge is enough. I have however seen devops jobs that seem to match my skillset - developer with a bit more. I hadn't really heard of the term until I saw the job ad.
As to DBA; I can't help but feeling that the OP hasn't worked with "real" DBA's. That's a whole different ballpark and I've yet to meet a sysadmin or developer who can make even a passable DBA.
I've always thought the hierarchy goes: DBA -> Ops -> Developers. With the last two really about equals.
When I think about expected earnings, I would say your hierarchy is correct.
Nothing you described is outside of the realm of what your typical linux admin does. I don't have to be a senior python dev to do my job, and I've managed 5500+ virtual machines by myself (puppet/chef, bash, some python, persistent data/object storage).
Agree with the author; just shoving more hats onto less people.
A DevOps can expect a bigger salary, while a company hiring one can expect way more productive candidates than if they asked for only ops.
Then dotcom happened and every kid with a Linux box in their bedroom put themselves about as a SA. And in the 10s people think SAs who code is an amazing new invention.
And back in the 60s, IBM had "systems programmers"... Same thing.
The big difference I see in devops is that people started taking system management seriously enough to do first-class development rather than an afterthought.
I still am, but the "DevOps Movement" is here to point out that this artificial dichotomy is considered harmful.
Oddly this is much more like the 'developers' of old. If you sat down at a workstation you needed to know how to be your own system administrator, and you needed to write code.
Automation has enabled a fairly new class of engineer which I think of as someone who has no idea how the pieces all fit together but they can assemble them with the help of a framework and a toolkit into useful products. They become experts at debugging the toolkit and framework but have little knowledge of how everything else actually works.
The problem with this new type of coder is that they can write syntactically correct impossible programs. I didn't understand that until I taught myself VHDL (a hardware description language). VHDL was the first "language" that I knew where you could write syntactically correct "code" which could not be synthesized into hardware. The language expressiveness exceeded the hardware's capabilities (and sometimes you would need to have a time machine). Imagine a computer language where 1/0 was a legitimate statement, not caught by the compiler, but always blew up in your executable.
So we have folks who can write code that is grossly inefficient or broken on "real" systems.
Google had started a program when I was there to have developers spend time in SRE (their DevOps organization) and it was to invest in them the understanding of what went on in the whole stack so they could write better products. The famous 'times every programmer should know' by Jeff Dean was another such tool. You cannot get too far away from the systems that are going to run your code if you want to write performant code.
In 2009, DevOps seemed like there was finally a reasonable answer to Taylorism. Engineers and Programmers and Hardware Technicians and Support Representatives were not cogs in a machine, but humans that could collaborate outside of rigid boundaries. Even at the lowest levels of the support organization, individual workers along the chain could now design their own tiny factories.
From there, it's just a matter of communicating tolerances properly up and down the chain. I am probably over-romanticising these notions, but it certainly felt exciting at the time. Not at all like the "fire your IT department" worldview it turned into.
... isn't that all languages?
Second, my only point was that the example given was a piss poor example of the difference between hardware and software. Obviously a bad example doesn't disprove the claim it's supposed to support.
There's been an awkward growing phase of the technology industry that has led to technicians that don't have any real understanding of the systems they maintain. Compare and contrast Robert DeNiro's character in Brazil with the repairmen he has to clean up after. We could be training those poor duct maintenance guys better.
Chuck points out that abstracting the Developer's work too far away from the system in question means the Developer doesn't really understand the system as a whole. Jeff refers to "purely development roles" and other "pure" roles that aren't necessarily natural boundaries.
The example of VHDL is not about hardware and software, but about learning that you didn't actually know something you thought you knew.
The repairmen in Brazil do not realize (or necessarily care) what they don't know about duct repair. The system allows them to function and even thrive, despite this false confidence in their understanding.
At one point at least, Google was investing in (metaphorically) having DeNiro cross-train those guys, instead of umm... Well, the rest of the movie.
The initial detail was that VHDL, unlike "software" languages, has very different consequences. Can you imagine a language where (1 / 0) wasn't defined away as a DIVERR, but otherwise managed to remain mostly self-consistent? Where something can be logically / syntactically coherent, but not physically possible?
And if that example didn't hit home for you, so it goes, but there was plenty of detail unrelated to the specific example that I thought was more important / interesting to discuss. :shrug:
And, contrary to the stated intentions, I've directly observed developers making crappy, band-aid fixes to ongoing production problems in the interest of "making the pages stop". This is the mindset when you are on call be being paged at all hours.
In theory, DevOps is supposed to put those that can best fix things closest to the problems, but in reality a slight separation from the firestorm of ops actually produces better, more thoughtful solutions in the long run.
The best balance is to have a first tier Ops on-call, 2nd tier engineering on-call, and any alerting issues get attention within 24 hours, moving to the front of the work-queue. But, indiscriminately assigning everyone "pager-duty" rotations leads to lower quality solutions in the end.
• It increases pager coverage, and reduces any one person's pager obligations. Simply having pager anticipation is a mental burden after a while.
• It creates a stronger incentive for response procedures: what are the expected obligations of response staff, what's considered sufficient effort, what's the escalation policy, who is expected to participate, what are consequences of failure to respond?
• Cross-training. Eng learns ops tasks, ops has a better opportunity for learning what eng is up to and deals with.
• It makes engineering more aware of the consequences of their actions: is insufficient defensive engineering causing outages (say, unlimited remote access to expensive operations), are alerts, notification mails, and/or monitoring/logging obscuring rather than revealing anomalous conditions? Are mechanisms for adjusting, repairing, updating, and/or restarting systems complex and/or failure prone themselves?
My experience at one site, where I was a recent staff member (and hence unfamiliar with policies, procedures, and capabilities), systems went down starting at 2am, I was unable to raise engineering or my manager, and the response the next staff meeting to my observation of this was pretty much "so what" did not endear me to the organization (I left it shortly afterward).
Note that what I'm calling for isn't for eng to be the sole group on pager duty, but for eng and ops to share that responsibility.
Within the right framework, keeping everyone on pager rotation can lead to much smoother operations, because everyone stays familiar with the system as a whole. This was going around recently, and captures the essence of the philosophy: http://catenary.wordpress.com/2011/04/19/naurs-programming-a...
At one place I worked we had a two-person support shop. We would claim time and again that this or that affected customers or made support hard. The devs would pick and choose what was fun to work on. I ended up leaving and the other guy went on a prearranged month-long vacation. Everyone else had to pick up support (~5 devs) for a month, and I'm told that they had so much trouble with the normal support load that development actually stopped for that month. Apparently when the other guy got back, they started listening a bit more to his concerns, having had a taste of what happens on the pointy end.
In a similar vein, there's a wine distributor where all employees spend their first week half on the phones and half in the packing department, to give everyone a feel of what the core function is and what customers complain about. The guy telling me said that everyone gets the treatment, except the new CEO, who got away with only doing a day rather than a whole week.
This is honestly why I've gone with PaaS - mostly Heroku - for several months now when deploying a new application. Why on Earth developers do anything other than working on the core features of their program I don't know. All of the things you need to set up - tesing pipeline, and containerized, automatic deployment, load balancing, databases - are now available as cloud services. There is absolutely no need for the developer to be doing administration and provisioning tasks at this point.
If you think you need to set up your own server infrastructure ask yourself one question: is there any specific technical requirement that my application has that can't be fulfilled by existing cloud services? If there isn't, and there probably won't be, you shouldn't be doing ops yourself, especially not in a startup setting where time is absolutely at a premium and you need to be spending all of it on making the best product you can make.
And before everyone tells me that PaaS is more expensive - it's only more expensive if your time is worth nothing. But your time isn't worth nothing - it's probably worth over $100/hr if you are a developer working in the United States. So Heroku ends up not being more expensive at all - especially not before you have to scale.
Core banking, batch processing, highly sensitive data stores? Probably not great candidates for public cloud consumption.
Web properties and services which don't rely on said functionality? Absolutely great candidates.
And the reality is, whether the IT guys like it or not, developers inside of these orgs are consuming cloud because getting a VM in a traditional sense takes forever (for good reason.)
As a result, we're seeing a shift in the industry where the idea of large corps / financial institutions / government are being pragmatic about the idea of a 'hybrid cloud.'
The reality is, adoption of cloud in the enterprise is growing, not shrinking.
Say you could be paid $100 p/h at some other company, but instead your start up is paying you $5 p/h because that is all it can currently afford.
Say you can save yourself a weeks work by going with one hosting option that is twice as expensive but the total additional cost would allow you to work for $5 p/h for an additional two weeks then that isn't a great trade because the "value" of that labor is both unknown and irrelevant at this point. It might be $0 p/h or it might be $10000+ p/h.
So even at $5/hr I can't justify doing this work myself at this point, since I can get the infrastructure for my minimum viable product going for free. I can and should spend my time focusing on developing the features.
Being an SA, what rankles me most is the attitude (unfortunately common in the industry) that "As a developer, I could naturally be the world's greatest Sysadmin, if only it weren't such a waste of my amazing talents."
I've suffered through enough three week deployment cycles because the prod environment is almost nothing like dev and everything is done manually. I think I know what "doing it wrong" looks like there.
In the old days, it was an Architect/Team Lead/Sr Developer who figured out how to distribute a product, then the maintenance and upkeep of the installation scripts was later handed to developers of “less” capabilities. But the architect still reviewed and was keep abreast of installation changes. An Architect/Team Lead/Sr Developer should setup the intial design, scripts etc for use as “DevOps”, but DevOps is not a new engineering discipline. The DevOps tools are trivial to understand (kinda like InstallShield’s VB scripts) and easy to master. However, it does require an engineering discipline. Kids that are used to pushing buttons in NetOps can’t suddenly become an experienced coder … they know little of classical software development.
I disagree with the author implying that you should hire a DevOps engineer to do this work so that coders can “code”. The economies of scale are way off, this is something a that a junior developer masters so that he can spend more time coding. I wouldn't recommend companies spend so much on an employee performing a task that is trivial at best.
No, this task is a Developer’s job, and as a Developer, you better know about them. I won’t argue that mastering these tools are not time consuming, but if you want to master development, these tools had better be on your roadmap. Software Engineering is a new discipline, what we need to know increases over time and mastering this field is getting harder and harder … kinda like natural selection.
DevOps is about developer empowerment. It's about creating the systems and tools that give developers more control over the operation of their applications. It's about removing Ops as a technical barrier between new code and live code.
At least, (as someone whom most would label as "DevOps") that's what it means to me.
That's not true. There are many scenarios where devs also understand ops and do both. It's getting rid of the "throw it over the wall" mentality and incorporating ops within dev itself IMO. And devs love to automate things so we have created a lot of tools for that.
The OP article is entirely missing the point, and you've set it straight. DevOps and the "Full Stack Developer" are entirely separate problems. DevOps can be specialized as well.
Like, I'm a full stack developer [e.g. I provision my own prod boxes, write the services that run on them] at my $DAY_JOB. I'm not seeing that as a bad thing unless it gets out of hand and I'm doing that for more than a small cluster of backend services.
One solution is to train up your ops guys in basic Ruby & Chef (for example)
It's anti-silo. Instead of group A building software in isolation and then tossing it over wall to group B to deploy. Those groups merge, cross-pollinate or at least communicate frequently. So, the group building software is away of the needs/issues of the group deploying/running the software and vice versa.
extend to other groups functions (QA, maintenance, sales).
It doesn't matter if one person does a bit of each role or if there are persons for each role. As long as the work closely together.
Sure seems that way at many places.
Like, I could use an Op for my stuff, but I also could automate "him" away.
When I was studying CS, there were a bunch of people, who didn't like to code, so they became Ops.
If Ops is now about programming, they can't even resort to this branch of CS...
The problem is that employers demand specialists, especially for senior positions. At the same time, once they've acquired an employee, they refuse to respect specialties from that point on. Machine learning expert? Sorry, but we need a ScrumDrone over at desk 21-B. Being a software engineer means resolving the fight between your job and your career, which is probably a big part of why this industry is so political.
Employers are remarkably inconsistent in this regard. They want sharp people who can interview like real computer scientists, but get in the way of their continuing sharpness (by assigning smart people to dumb work) as soon as they're on board.
The insight that the OP has is that employers over-hire for crappy work, and he's completely right, but DevOps didn't do it.
5-10 years ago a full-stack developer was a very meaningful distinction. Today, every hacker wannabe Uber driver that went to a dev bootcamp for 3 months calls themselves a full-stack developer. "DevOps" avoids this fate only because the subject matter is slightly heavier and harder to fake.
That's when DevOps gets really helpful and valuable. But if you haven't worked in environments like that, you have no idea what it's like.
Yes, we had automation in the '90s — I wrote quite a bit of it myself! — but the landscape has drastically improved. For one thing, the industry is now embracing it, and with the embrace, a name. The name is not being "sold" in any way that I can see — no one is getting rich by bandying around the buzzword. It's not being sold by anyone I'm aware of as some kind of silver bullet, and anyone who believes in an IT panacea deserves what he gets. It is however being used to sell an idea, that automation in the '90s and before was a good thing, and that we should probably do more of it. DevOps means more than just automation, and in large part, these are also improvements in the industry. We're better now, partly because we have to be.
For that matter, the cloud is just a name that describes the commodification of computing resources (whether that be actual compute, storage, whatever). Yes, yes, the marketing blowhards of the world have misused and bastardized the word, but that doesn't mean they've ruined it, or that it never meant anything.
I have no idea why you think Chef and Puppet are broken piles of ruby scripts, but for the record, they're free. Also, in neither case is anyone implying or saying that they invented automation. Having a nice framework to use is a definite improvement, though.
You have general contractors and then you have subs that work under them. A general contractor is a jack of all trades, master of none. Exactly what a full stack developer is.
This isn't the end of specialization. It's the beginning of project management steered by developers who intimately understand all of the work involved, even if they aren't as competent as the specialists.
Having a team consist of all full-stack developers is just stupid. Having a full-stack developer as the head on a project, with specialists on the team, is a great idea.
A software lead is one a type generalist, as they have to manage resources that can X, Y and Z. The generalist position described in this article is someone who can and will dive in and do X, Y and Z themselves. Very different roles.
It's the beginning of project management steered by developers who intimately understand all of the work involved, even if they aren't as competent as the specialists.
People with a wide breadth of general development knowledge, once employed as developers but now managing development? I wouldn't exactly call that a new idea. Or are you talking management that also develops across all components of the software? If management takes up so little time of a single developer's time, then you're just using different words to describe the small team constraints that the article does.
In practice, though, it seems like the minimum bar for entry becomes "full-stack experience" and then understaffing, because why bother hiring when these people can do everything anyways?
That's how the HR and operations folks screw it up.
I go back a few years, to an old, waterfall-like job. I was handed work by an analyst, that was handed a task to analyze by an engagement lead, who might at some point talk to someone using the application. The work was always handed out on time, but the product often failed, not because it was buggy, but because nobody actually had much of an idea of what we were really trying to solve.
So us developers got much work done, but the work didn't actually solve real problems: The force is applied to the wrong vector. Then the product fails, and the blame game begins: Changes are too expensive, because the developers didn't know what the real invariants are. Queries are slow, because the database architect wasn't told about the questions that the database had to answer. The business analysis just wrote documents. It was all a big catastrophe.
That company moved to Scrum, the terrible way: Here, have this self organizing team full of specialists that don't know anything outside of their domain. They are still failing to this day, but now they blame each other in retrospectives.
So I'd much rather be stuck coding less, but then being aware that my code is actually solving a problem for someone, than just writing castles in the sky, because everything I've been told about what my userbase needs comes from a game of telephone.
I diagnosed a few problems over the years that arose as apparent issues with a web application but that I gradually narrowed down to things like network issues, or kernel bugs, or system misconfiguration, or database issues etc. Modern stacks are very complicated and the interactions can get really messy, it is close impossible for someone who doesn't understand the whole thing to find issues that aren't neatly isolated. I perfectly know that I do not have the full qualifications of a sys-admin proper, and would not like to do a sys-admin job full time, but in those particular cases a pure sys-admin would not (and often actually could not) find those issues. As an example, I can remember many situations where the application showed different behaviour depending on which application server you hit, and typically both "pure" developers and "pure" sys-admins were having a hard time finding the issue.
Good sys-admins anyway have to learn, at least, C programming, shell scripting, and network protocols and programming, so it's should not be a big deal to add some Rails/Django/Node to their skillset. Good developers anyway have to know things about hardware, networks, protocols and so forth. You do want to have people that are specialized in one or the other area and focus on it on a day to day basis, but you also do want to have people that can understand a particular aspect of the system top to bottom when such a need occurs, and it does happen quite often.
I'm all for a tighter integration between Ops and Devs, and infrastructure-as-code can help bridge that gap, but I don't know that doing each others jobs is the solution.
At the plus side, you have quite a big deal of control over most of those variables.
Having developers be 'full stack' imho reduces the amount of "works on my machine". How would a developer test the software he/she is developing on if she can't at least get close to a production environment.
Automated provisioning is just one of the usual 'devops' things that I can't imagine how a proper software engineering process would work without.
I would say that at least 20% of the people I graduated with can create software that works mostly ok when they hit the little green "run" icon in Eclipse. They were however incapable of figuring out why their jar file doesn't work in tomcat on a linux server somewhere.
Usually it was because they're using a local database with root credentials instead of a remote Database with multiple users, they have some file stashed away somewhere in their classpath, they have some binary installed in $PATH that makes the whole thing work.
I think just wanting to be a developer and not know about the stack that your application runs on is like being a painter but refusing to buy paint because you can't see what going to the store has to do with painting.
That limits the crazy stuff like esoteric package function for the database, crazy port opening on the machines, esoteric daemons, or tools that can only be deployed from sources. Because the guy deciding that will be the one updating and debugging the VM creations script (for 2 different cloud provider).
But at the same times, they have access to all the tools they want, they can evaluate if something is best done on the admin or development side, they use mostly the same tools, code conventions and languages for administration and product development, and they have an immediate feedback sur how much logging to send from the application.
Not necessarily, it can also be worsened. I think Conway's Law is especially appropriate here: "[O]rganizations which design systems ... are constrained to produce designs which are copies of the communication structures of these organizations."
When everybody shares IT responsibilities "because they can", it leads to each of those developers to run their own little customized fiefdom, with subtle (or not-so-subtle) version mismatches, mysterious lurking cronjobs, and mystery-scripts that do per-app housekeeping.
All too often I see feature developers say "well, I have an operations team, I'll let them figure it out," and they (a) never leverage said operations team for advice during development, (b) don't consider operational concerns such as sharding, deployment, logging, and monitoring at all during development, (c) file a ticket against operations with three weeks to go until their deadline to perform all of those things, and (d) call for rolling operational heads when their service does not perform to their expectations (using the author's "totem pole" as their rationale). As an operations engineer, I can count way too many fucking times I've been on the other end of that from developers with attitudes like yours. It is the absolute worst part of my job.
Such people exist and call themselves developers? Ah man.
First, DevOps has degenerated into a meaningless buzzword to rival "Agile", despite the good ideas and good intentions. Every day, I have recruiters looking for "DevOps". A couple of years ago, they'd never heard the word.
Second, DevOps is actually getting strongly biased toward Ops, often to the exclusion of Dev. In the eyes of recruiters and much of the industry, it's become synonymous with "Chef/Puppet/Ansible automation", a set of automation tools. That's stupid.
Third, and this is what matters to me... DevOps is (or was meant to be/should be) more about organizational structure than skills. As the author points out here, specialization is good and necessary. But specialization comes with bureaucratic compartmentalization that makes working across org boundaries very difficult. When you have to climb four or five (or more) layers up the org chart to find common management for both the dev and ops sides of a project, then the dev team has no authority over and very little way to communicate with ops, and vice versa. For most large organizations, the dev/ops separation is necessary - developers get locked out of production systems to keep them from legal exposure to customer data (HIPAA, PII, etc), and to keep them from accidentally or intentionally altering production in a way that it might break.
Read Gene Kim's excellent quasi-fiction book, The Phoenix Project. It covers a lot of the issues of DevOps as fixing communication patterns in large organizations. You'll see how little of it is about tooling or "full-stack", and how much is about clearing bureaucratic obstacles to effective communication.
I think this is going to happen to more and more careers. Already a profession like surgery or piloting a modern airliner is starting to require some of the skills we think of as programming. Software is eating the world - that doesn't make domain expertise irrelevant, but it means you need people with domain expertise and programming skills. That applies to non-programming roles in the software industry just as it applies to other industries.
Yes. Chefs also do shopping, menu planning, prep, hiring, firing, marketing, and schmoozing with patrons. Source: I know a chef.
You have to ask whether you're talking about a senior executive ("Chef de Cuisine", "Head Chef", "Executive Chef"), a subject matter expert or mid-level professional ("Saucier", "Pastry Chef", "Line Cook"), a junior person ("Prep Chef"), or a low man on the totem pole ("Busboy", "Dishwasher").
"Chef" is a very vague term when applied to large restaurants. Obviously, in a little family owned place, a lot of these roles would collapse.
I don't think most developers have the capability to be sysadmins or QA. Vice versa, too, quite often. Joe developer ain't that special.
Devops is about taking moving the infrastructure into its own configuration-managed artifact, taking lessons from programming and computer science, and coming out with its own engineering rigor.
If you want your devs to operate builds/infrastructure/etc/etc, that's fine, but devops that ain't. That's called "many hats".
- systems is about thinking how different things fit together
- development is about building something
- quality assurance is about figuring out how to blow something something up
The thing is, I know technical people that I wouldn't trust to do one of the above, let alone all three at once. And I know people who can rock it in multiple disciplines. When it all comes down to it, focusing on the quality of the people you work with and helping them thrive is a solid plan.
No, but they tend to think they can.
It has nothing to do with limited startup resources, and everything to do with managing externalities.
Specifically, developers have an enormous amount of control over the stability and deployability of their software: technical decisions made at almost all levels of the stack directly and significantly impact the costs of QA and Operations.
The people best suited to automating deployment and ensuring code quality are the people writing the code.
If you entirely externalize the costs of those two things , natural human laziness takes over, and developers punt bugs and deployment costs to the external QA and operations teams, ballooning the overall cost to the company.
The burnout aside, there is a plus to someone being proficient at many related tasks; having a somewhat in-depth knowledge of how all these technologies come together. The point is not all jobs require the best, most expert techniques. As in the case of the jack-of-all-trades carpenter, as long as he knows when to call the specialist, he is still getting the jobs, as am I.
This isn't counter to specialization -- in a big organization, people are certainly still going to specialize. But the "DBA" equivalent people are just programmers who have fresh expertise on the storage layer, and the "QA" people are just programmers who have expertise on the automated build and test systems.
The dentist analogy doesn't hold in software. A dentist handling secretarial work is just an expensive waste of time, due to comparative advantage. But a programmer replacing secretarial work with automation often reaps big long-term dividends.
However, I think as with most things that involve computational thinking and automation, this is not a zero-sum game. A developer who can apply deterministic, testable processes to server-ops may be able to reap an adequate amount of benefit for significantly lower cost than a specialized sysadmin. In addition, the developer is augmenting his/her own skills in the process. Yes, that dev was not able to focus all of their time on...whatever part of the stack they are meant to specialize in...on the other hand, the time spent studying dev ops is not necessarily a sunk cost.
For my own part, I've tried to stay away from sys-admin as much as possible...but when I've been pushed into it, I've gotten something out of it beyond just getting the damn server up. For example, better familiarity with UNIX tools and the importance of "text-as-an-interface"...which does apply to high-level web development...nevermind the efficiency you gain by being able to stay in the shell when most appropriate (rather than, say, figure out how to wrangle server commands in a brittle capistrano script).
But hell, even the end product itself, just being able to deploy a server with some confidence...is kind of empowering. For me, it opens up new ways to run scripts and jobs...It sounds dumb and maybe it's just the way my brain poorly functions, but the concepts of server-oriented architecture become so much clearer when you can spin up different machines to play with and experiment with delegation.
I don't see "DevOps" as a way to replace some roles - but as a way to make everyone work better together. Instead of living each in their own bubble (and in my - pretty limited I admit - experience it always) everyone has to know, at least a little, what someone else does. It really helps everyone at the end of the day.
And the developer can keep coding without me screaming at him because he placed the database connection string in a configuration file that sits inside a .jar that sits inside a .war and so on.
I'm just a guilty as anyone else in try to write code to fix a problem that's better handled by existing infrastructure and servers. I worked on a project to deliver invoices to customers in downloadable form. In the end pretty much very thing was thrown out and I just need to write the authentication part, because the sys admins pointed out that the existing StringRay boxes could handle everything else (http://www.riverbed.com/products-solutions/products/applicat...).
It's not that it doesn't make sense to have dedicated developers and systems administrators, as the developer you just need to know enough to be able to talk to and understand the admins position and thoughts.
"devops" is having developers sitting with, understanding, architecting, and in the end programming solutions to what were traditionally ops/sysadmin problems. and operators sitting with, understanding and participating in service architecture, teaching about livesite realities, coding where possible, and appropriately buffering devs from noise on the livesite.
the unfortunate thing is that many companies swing the pendulum too far one way or the other. neither all devs nor traditional dev + IT/ops orgs are the best way to build a great product and run a world class service.
However, it's obviously better to separate matters and offload management/administration tasks to a separate team/person. Thus, a good developer in a good company (which has a separate sysadmin roles) indeed can't really replace good sysadmin because the latter has niche practical knowledge on handling various situations (especially emergencies) quickly.
Nonetheless, one can be both a good developer and a good sysadmin at the same time.
I suspect in most mid and large sized companies that the top 10% of developers are creating more value than the next 50%, and that below that, the value could easily be negative.
Instead, about 40% what was called 'sys admins' were pushed to become devops. The 'sys admin who knew cfengine' became a 'devops person who knew ansible'. Deploys and cloud APIs just became another thing to automate.
The bottom 60% - the shit ones who got paid 120,000GBP to copy paste commands they didn't understand from word documents into Solaris 8 boxes in 2010 because they couldn't actually automate anything - left the industry.
Most engineers simply don't have a meaningful understanding of the internals of large-scale environments.
I wish. They are still here and still copy+pasting commands they don't understand. The entire linux world is still dominated by these people. The whole "howto" culture is still very much alive and kicking.
I might only be slightly better than someone who's new to system administration only because I've written system-level code and understand operating systems and things of that nature.
However a good system administrator understands the entire architecture from a holistic point of view. They know the compiler switches to use, the run-time switches to tweak, the security implications of various configurations and all of the other details it takes to keep a cluster secure.
I often work well with a good system administrator to debug and optimize workloads due to the overlap in our skills. I find this to be the optimal relationship.
Learning and practicing system administration takes away from my ability to learn and be a better programmer (and the opposite is true as well). I don't know about most people but I find I can't be good at both. And I know which one I'd rather be better at (programming).
I don't think the author has hit the nail on the head but I agree that effective teams can't expect one person to manage an entire application from code to managing a secure deployment.
- Encouraging collaboration between your Dev, Ops, and QA teams, with some cross-training so they can work together better
- Merging those teams under the same manager to try to improve that collaboration
- Making your developers responsible for all those roles, and never hiring a dedicated sysadmin or QA engineer
I personally think any of those is fine. Startups will err toward having fewer people and all of them be developers, while in a larger company it probably makes sense to specialize more and make "DevOps" mean close collaboration between those teams.
Of course, I've also seen "DevOps" as a job title for what would have previously been a "system administrator" or "site reliability engineer", and I have much less patience for that. :) Occasionally I see a job posting for a role that is actually dev + ops, but most often a "DevOps" posting means "we need a sysadmin, but we don't think sysadmins are cool enough to work here."
Some of our goals included:
- Building the continuous integration/delivery pipeline
- Moving codebases from one source control system to another
- Creating programs/systems to automate tagging of builds
- Automating the deployment processes of multiple applications onto non-production servers
- Implementing and maintaining the functional testing frameworks and server grids
- New development happened sporadically; day-to-day work was a mixture of maintenance development and admin work
- Culture. They started with a small team, and never grew it. Having more people didn't fit with the way the company saw itself.
- Difficulty hiring specialists. Various reasons for it, but still valid.
At another company I worked, there was a lot of "integration development", where your time was spent connecting various internal and external systems together, software-wise, developing tools that support systems work (i.e., tools for sysadmins), and developing other tools that are for end-users, but have a heavy systems component (management software for DNS, for example.) That meant understanding each part of the stack from both a development and system perspective. Another is interest level. A few of us were full-stack developers because we were studied more than just development in our free time, and we took that with us to work. This wound up benefiting everyone. This also led to us being the go-to people (that is, the top level of internal support) for both more specialist internal developers and sysadmins, as we had deep knowledge of the internal systems from the bottom to the top the stack, and the knowledge and experience to explain and troubleshoot problems to people in those other roles.
The author is correct in that this may be more /common/ at startups (the previous startup I worked at did in fact operate as the post describes), and is sometimes done out of necessity. It is by no means limited to those environments, however.
Edit: I'd also separate DevOps from full-stack engineer. They sound like the same thing, and if you squint from far enough away, they look like the same thing. The terminology may be fluid, but I think (as some other comments state), that DevOps is more centered around "coding for systems automation", whereas "full-stack engineering" is a much more general term which can encompass a variety of different types of tasks in different environments with varying levels of knowledge/experience in the different parts of the stack/tools.
That said, I don't think that the increased prevalence of DevOps is bad. And I don't think it means "everyone is doing everything" either. It's a new role that is borrowing elements from both development and operations. Not one person doing both roles.
Developers take on the new responsibilities of being able to independently deploy their code, instrument and monitor stability and own test/QA.
Platform Engineering is about building a robust infrastructure and the tooling needed for Developers to handle the new responsibilities. This includes packaging, monitoring, deployment, AB testing, etc.
Site Reliability Engineering is about dealing with fires outside of the codebase. Hardware failures, network connectivity issues, etc.
I don't think any of these roles becomes a "Jack of all trades, master at none" situation. It does, however, cut out some of the more typical engineering roles. While developers just took on additional responsibilities, QA engineers and traditional Ops are forced to repurpose their skill set.
DevOps is not about startups, DevOps is about avoiding the pitfalls of big companies who completely fail and leave all of their employees jobless by focusing on all of the wrong decisions and initiatives.
It's about outlawing cowboy coding and other bad habits that people pick up as hobbyists, and intertwining business and technical objectives reasonably.
Why is a full-stack developer important? Why is eroding the difference in responsibility between Dev, Ops, and QA important? Because traditionally along these boundaries have been opportunities for individuals to absolve themselves of responsibility. More than anything, DevOps is about not living in that world anymore.
Some people won't survive outside that world. Those who want to will read "The Phoenix Project" by Gene Kim.
1. The TCP/IP Guide: A Comprehensive, Illustrated Internet Protocols Reference. http://www.amazon.com/The-TCP-Guide-Comprehensive-Illustrate...
2. Advanced Programming in the UNIX Environment http://www.amazon.com/Programming-Environment-Addison-Wesley...
3. A systems programming language- I choose golang.
5. SSH, The Secure Shell: The Definitive Guide: The Definitive Guide http://www.amazon.com/SSH-Secure-Shell-Definitive-Guide-eboo...
Also, not sure what you're getting into but usually for sysadmin for scripting you need to know one of: Bash|python|Perl|Ruby
I am a manager/developer/architect at a relatively large software company, and we have to task our developers with devops-type tasks constantly. Not because we want our developers spending time outside of coding, but because for lack of ability to hire the competency needed.
As you stated, good developers can generally perform these tasks so when you have nobody lower to perform them they become a weight on the developers' shoulders.
No it isn't necessarily fair, and yes, I believe in the future specialization will come back as the education system starts to realize there are many jobs in tech, not just a Comp Sci degree jobs.
This is known as the "odd man out" syndrome. I currently work in a medium sized company who are doing a huge ERP switch over. I'm a front-end developer by trade, but know .Net as well. One part of the contract stated our company needed to have X amount of company resources (people) to have on the project.
The downside is I hate my job now and am actively looking to get out of here. They told me recently after the release, I'll be one of the ongoing "resources" to help manage post-release defects.
So I agree on your last point as well. It's not fair and unfortunately, it's a no win situation for the developers. If I do a shitty job as a JDE developer, they get pissed and might evenutally fire me. If I do a good job, then I get tasked with all kinds of stuff I have neither the want or desire to do.
Six months in and I hate working in JDE but all the contractors think after two weeks I should be a pro with it since I'm a "developer".
I'm not understanding this, you can deploy with Puppet, Chef, Salt, Ansible, Vagrant and Docker. With Vagrant you can deploy a bare image and use Chef (or one of the others) or you can just deploy a fully setup box file (like with Docker).
With distributed systems becoming the norm rather than the exception, developers will need to understand how and where their code runs in production (and how it gets there) to be able to debug issues or write better behaving code.
(sorry if the q is off topic, I don't really understand what OP is ranting about with the devops problems, so I'm referring to the only part of the article that makes any sense to me, that about the full-stack devs...)
I can relate to the downsides pretty well - I'm the only developer in my group and my job is mostly to develop web apps, but the IT side doesn't have much knowledge of modern tools - they live in the era 'just use Drupal and Apache' so I'm often the one who ends up having to figure out the deployment of the applications I work on (and also help with random problems from their OTB apps) and such.
To be honest, I don't mind when it's DB stuff because I'm pretty comfortable with it and have plenty of background with various SQL DBs, and it's not a time black hole, but when it comes to configuring servers and deployment I hate having to deal with the DevOps because there are so many pieces I never have the time to really become comfortable with them all and I feel very inefficient. Accomplishing something doesn't always take long itself, but it can require spending a day of reading wikis and documentation to accomplish something simple when you've got a lot of moving parts. And the worst part is that you have to deal with the DevOps bits so infrequently it's like you have to relearn them each time.
I have worked with DBAs who had PhDs and could have still done development, but they moved past that to concentrate on schema development, scaling, etc. Toss into the mix modern programs of master data development inside organizations and people who are characterized as DBAs have a very sophisticated role.
Also, for small projects, devops makes all the sense in the world to me. Deeply understanding how an entire system works is valuable.
It means that no part of the stack goes not understood, or .. all parts of the stack should be understood and controllable by the developer.
Guess what - this doesn't produce 'worse developers' .. it produces better stacks. The fact is that the fracture and delineation between the cultures of code, rather than the actual code itself, is the true danger. Getting 'the db guy' to talk to 'the front-end guy' is a posers game. Get rid of it.
Instead, get your guys to move across the tree of responsibility that a full-stack approach requires. In truly professional development, there is always going to be new things to learn and new things to use to manipulate the machines - this is turned to an advantage in the full-stack approach since it requires an adherence to a real policy: you just don't care about 'the culture of the tech', you read the docs, you write the code, you read a hell of a lot of code, and you don't really put limits on what you can and cannot understand; those limits are instead expressed in working code, at any layer of the stack. The 'cultural excuses' for why things are borked 'over there' are no longer relevant in this approach; if you're a real full-stack guy, you'll get along - source or no source, but hopefully: mostly always with the source.
It is a political approach, but it works - especially in industry. There are a few other principle-based disciplines in the world where an 'all-embracing' privilege exists, in this case we are lucky that computers, as grand engines of word and significance, are a form of literature. Study well, and study all .. to the end!
Waterfall is not just seen as broken, it was always broken.
I prefer to think of it as the consequence of software development being tacked onto most other existing processes, without regards for the practical needs of software development.
- (pure) Specialists often don't understand how their decisions affect other systems (and middle management or communication isn't always a solution)
- (pure) Specialists tie you to a particular technology when in reality you may need to evolve to use other technologies.
- If you need a bunch of different specialists to get something simple done (perhaps something you don't do all the time so don't have a process in place), just because they are siloed, it's a lot more complex and usually ends up badly designed (because it's harder to be iterative across teams). Generalists can get simple things done that require different skill sets to accomplish.
I'll agree with your second point, to some extent. Generalists are rarely tied to any one technology and therefore can be very good at getting your organization to use the right tool (avoiding the common "hammer-nail" problem). However, just as frequently, I see generalists picking the wrong tool for the job, because they again aren't intimately familiar enough with the tools already at their disposal to understand all their capabilities or be able to make an informed decision about whether the gains of the new tool are worth the added complexity of introducing another layer to their stack. And, of course, nobody feels the pain of adding extra layers to the stack quite like DevOps do.
I'm not sure what to make of your third point. Isolated processes talking to each other is just a different strategy from a monolithic design. There are advantages and disadvantages to each. It's unclear to me that monolithic means "better designed," and in fact there are good security arguments to the contrary. But maybe I'm misreading what you're saying.
It's much about applying the same processes you would apply in development to Ops. For example committing changes into version control and only using that, not live patching things, much like you wouldn't live edit a website.
Also being able to spin up new servers based on a config and not requiring manual config to get it going. Automation alone does not get rid of 'snowflake servers' http://martinfowler.com/bliki/SnowflakeServer.html
Also, it can be about letting developers get the exact same environment for development/testing at no additional time cost - which in turn makes it more like that code changes can go live without problems or delays.
You still have specializations - SDE's, systems engineers, DBA's, etc. However, if you write code and it ends up in production, you are responsible for the proper function of that code in production. As a friend of mine put it in terms of developers who don't want to be on-call: 'what, you don't trust the quality of your code?'
DevOps is simply a nicer way of just saying, "own your damn code." The corollary to this is that the organization must help you in getting to that state where you can effectively own your code - this means collaboration (so that you build maintainable systems) and building tools that enable fairly frictionless code ownership.
The anti-DevOps developers don't understand how databases work so they want an ORM to make it easy for them, and they don't know how to configure a web server so they want PaaS solutions to let them do 1-click deploys, and system command prompts are scary to them. Frankly such developers just plain suck. They don't like DevOps because they don't have the skills to be DevOps.
In general a Full Stack DevOps oriented approach always tends to be more efficient. You have less monolithic hard to maintain applications because you force the teams to be small and agile. People will have their specialties (operations, backend, frontend, etc.) but still remain generalized enough to have an idea of the big picture. If your application has issues where the frontend developer doesn't know the general idea of how Varnish and Nginx in your stack are setup then perhaps your application is too big and complex.
When I think of DevOps, I don't think of having everyone know everything. Ops staff have to know enough code to write deployment automation scripts and dev staff need to know enough system administration to step up and help when the monitoring or deployment automation breaks.
It's meant to be a partnership to maintain a system rather than the old practice of throwing code over the wall. It really harms morale to have the developers all enjoying wonderful weekends while ops is on red alert because app changes they don't understand broke everything in production.
The bad side is that doing this DevOps role across multiple projects at the same time can lead to burnout, and I think I came close to that in the last few months.
The good side is that I've learnt a great deal about how to architect and deploy distributed web systems, how to do end-to-end testing, and how to effectively run the ops side of the business.
It's a mixed bag, and the burnout is the worst aspect of it, as well as the case where people are forced into situations where they are way in over their head.
If not, then aren't you being taken advantage of?
I'm curious to know how specialists negotiate for pay or write a resume if that isn't their specialty.
To suggest otherwise shows a complete lack of understanding of the nuance of those roles.
As for suggesting that "DevOps" is killing the developer - the only thing "DevOps" is doing is polluting our common language with a term that doesn't actually mean anything concrete. It's perfect consultant speak.
DevOps is stupid because it fractures expertise and makes it more difficult to get work done. By splitting up roles you get more domain-specific knowledge, have more time to work on a single problem, and provide support for your co-workers who also have different specific roles. I would much prefer to work with specialists than generalists.
You hate xyz? OK, but apparently xyz has enough merit to get the attention of quite a few people, so let's identify the problem areas and make xyz better rather than resorting to hyperbole and melodrama.
You can't have a fruitful discussion when everybody uses it differently.
Maybe that wasn't true "DevOps"?
But they cost much more as well. Following that logic, it would be in the interest of hospitals to make "full-stack-doctors" clean toilets.
... because DevOps is just a buzzword at this point.
He makes some good points but he misses the value of needing fewer people to accomplish the same thing.
The issue is that employers are horribly inconsistent. They demand specialism in hiring, but refuse to respect specialties once they've pulled people in. Thus, you end up having to interview like a real computer scientist, only to find that most of the work is mind-numbing for a serious programmer, but that there's no one around at-level for it because "we only hire A players".
DevOps didn't do this. The problem is the industry, not one concept.