Ask HN: Do you (developers) manage your own servers?

I used to, but ain't nobody got time for that. Unless it's a three-people shop, in come the devops (developer time is far more expensive than devops engineer time, at least around here, so making a dev wear a devop hat is wasteful on several levels - comparative advantage and all that).

Not sure if you had a subset of audience in mind, but in the "Enterprise world" (However loosely we define it:), it's extremely uncommon. The roles are highly delineated, and developers are usually insulated from underlying platform and technology. I've worked public sector, finance, retail, manufacturing, and have rarely seen a developer even interested in anything beneath a high-level programming language and SQL.

For my side project, I manage my own VPS servers. I had to do all the install and security for them.

I'm a Freelancer and run my own business. I do most of my server administration.

I am a freelancer, I manage about a dozen client servers and couple of my own VPSes.

As a developer with personal projects, I do all system admin stuff for my side projects.

I have to manage my servers (dedicated and VPS both) myself as I don't have money/resources to hire a dedicated devops person.

One of the things I wonder, in reading the comments... I wonder if there would be a viable model for someone to manage the devs servers for them.

I'd love to do this, but I wonder what a good pricepoint would be? (so that it's beneficial to both parties)

Yes, but we are a small group with modest needs.

Yes, currently approx 80 servers, fully automated, they mostly manage themselves.

The only involvement I have is the weekly security patches, I review the upgrade list and then confirm it myself before it is deployed. This way I know if any services will automatically restart, which can have some implications, this is rare though

Yes, as a solo-preneur or on a small team I've always managed or shared the responsibility. It can be fun, but it is amazing how much time devops takes away from product delivery. I can spend hours getting logging setup, or figuring out Azure resource group template files...

Thats why I've been gravitating to Openshift/Heroku/Dokku style push to deploy. And for the things that don't deploy nicely there I can just spin up a docker-compose.yml file and deploy that way.

I'm only a few months into this new workflow, and it is paying off. New ideas and projects are easy provisioned and even bootstrapped with a few commands. Sure I spend more in compute resources, but I'm glad to have my time back.

No, I don't handle the whole server, but I think I do more than many. I write all of the HTML, CSS, JavaScript, PHP, and SQL, including the DDL. Further, I must customize all of the config files for Apache, PHP, and PostgreSQL from stock.

There are other teams that handle: installation and upkeep of the Linux OS, back-ups, virtual host, physical host, and of course, the network.

I work at a large company whose official standards are Microsoft SQL, Microsoft IIS, and even Microsoft Team Foundation Version Control. I get away with using an open-source stack because it was established long ago, it supports several important web applications, and my stuff is down much less often than other teams' stuff, keeping me out of sight, out of mind, most of the time.

Yes. I manage the servers for a small startup I'm working on, we sell our software to the enterprise and have to do a custom deployment for each one of them which is very time consuming. I'm considering installing cloud foundry and doing all the deployment through that

I work at an early-stage startup where we all do a bit of devops, but I end up doing most of the configuration.

I think I can answer this question partially from a slightly different stand point.

We host (the fully managed kind) several clients' projects (small and medium). It isn't atypical to have some clients hire us for server "management" as insurance. Meaning we only dive in when something awry happens to their stack and they need that jack-of-all-trade tech who can pinpoint the culprit quickly. ~$30/mo as insurance is peanuts compared to potential revenue gaps.

Others want a turnkey solution where we document the details of their stack and we manage the ops part in coordination with the dev. I think this is where the most value is captured by our clients.

There are other interesting outlier cases but it's likely OT :)

At my job? No.

My sense is that it depends on team size -- when you are big enough that there's full time work in DevOps, I think you should have a specialist.

We have a mixture of dev-managed, and ops-managed servers here. However, we are exclusively moving to a model where dev create chef scripts for servers, but ops manage them.

I am not a full time dev, but do write little scripts and web apps to help out my team members. I have a server that I have to support for this (no external support) and I actually like it that way.

I do, and I actually enjoy it. A lot of management today is controlled by scripts (at least in my env).

Not fully a developer nowadays, but I either self manage or use cloud services.

Lately I have been thinking about hiring a freelancer to optimize some servers, but other than that I control the servers for my business. I should probably pay someone to help me since I use my servers for tracking/redirecting my campaigns, and if a server goes down I am just losing 100% of money.