Hacker News Reader

Comment by peterwwillis | Hacker News Reader

Who here would like to contribute their actual organization's practices to a wiki just dedicated to SRE stuff?

Rather than just talk about how we're supposed to do this stuff, we can share how we actually do it. Pick some recipes (procedures, configs, documentation, tooling, workflows, etc), learn from previous teams, and go to town. Contribute back the things you've done/changed/learned to the wiki.

I bought the domain sre.pizza because it was funny and easy to remember. I can set up a wiki there this weekend...

View on HN

There is a story about a motorist stopping in the Irish countryside and asking for directions ... "I would nae start from here if I were you" being the reply.

companies SRE efforts range from world leading to fiscally irresponsible- and revealing the latter to the world will be actionable - against the "whistle blower" and yourself (publisher).

Wikileaks looks like it does to protect against that. (partially)

So you can either look like wikileaks, and have a genuine survey of current state of play, or you can have a "this is how we did awesome SRE at the tiny portion of the big co I was employed by" talks.

I don't have an answer if you are in the irish countryside and want to start from somewhere else.

That would be a question worth answering

If I had to make a wild guess, I'd say companies that are seriously willing to engage in real SRE are probably also open enough to have a "How we did it" chapter written. At least in the Fortune 500 world where I live, companies aren't into pretending they don't suck, generally.

And the companies that are likely to be jerks about it are unlikely to be doing interesting SRE work. I'd like stories of hard-won success (or cautionary tales of failure), not just ranting about how much Company X sucks and how broken they are.

Exactly. Most companies are willing to advertise their wins and lessons learned because it means they grew. If they don't have wins yet, they can use the wiki to work on their process and eventually contribute back what works. And I'm sure there's lots of existing companies who have gone through all of that (several of which are in the SRE workbook)

Fortune 500 companies are vast agglomerations of different companies, fiefdoms, practises and teams. The same company could hold world leading teams and teams making every mistake going.

And the legal department will be jerks if the latter get the spotlight.

Please launch it - I hope it will do well.

There is something like that for security started by F500 companies: https://www.bsimm.com/

Their report breaks down some stuff on what should be done on what maturity level and then it shows data on how many actually are doing it.

Anyone have a link to the actual document that's hidden behind the harvesting form?

edit: a web search picked up https://www.bsimm.com/content/dam/bsimm/reports/bsimm8.pdf, I'd guess it's this one.

As someone interested in learning about SRE stuff I think this is a great idea. Any particular resources or areas you believe a beginner should focus on to build solid foundations? Thank you.

I think the SRE book and the workflow are good guides for some specific aspects/examples of SRE work. There's also tons of books on industry best practice and standards (none of which I can recommend specifically, maybe someone else here can?) ranging through things like ITIL[1], risk management[2], IT management[3], OAM[4], operations management[5], network and service management[6], etc. These are more abstract than "how do I set up monitoring", but they give you an idea of the breadth of what a large-scale organization needs to run high availability, high performance services at scale.

[1] https://en.wikipedia.org/wiki/ITIL [2] https://en.wikipedia.org/wiki/IT_risk_management [3] https://en.wikipedia.org/wiki/Information_technology_managem... [4] https://en.wikipedia.org/wiki/Operations,_administration_and... [5] https://en.wikipedia.org/wiki/Operations_management [6] https://en.wikipedia.org/wiki/Network_and_service_management...

Thank you for the thorough response, I appreciate it very much.

I'm more than willing to contribute. I've been looking into such an organization for a while now, I'd be happy to contribute back the "problems" and incremental "solutions" I'd like to test out. In my case it's a bit different, i'd love to see SRE being implemented at a large scale for packaged software where the customer operates the bought product on his own.

A lot of organisations prohibit sharing internal practices — or at least make the process really hard by requiring multiple approvals.

Reply on news.ycombinator.com