I promised more lessons from 2018, well here's one: you can be a good engineering manager for infra teams, and not know fuckall beans about managing a team of *product* engineers.
i learned so stupid much just from watching @eanakashima roll (she's an extremely strong product eng manager, who we hired out of sheer dumb luck) as well as @samstokes, @cyen, @ellenchisa & other folks who have thought deeply about product eng process.
lord knows i sure hadn't.
the thing about infra is how much it is relentlessly driven by what *must* be done.
oh there's creative work to be done, bet your ass, but it's your job to make sure there's as little as possible. and key fact: your consumers are usually other machines. (thank ghods)
there's still prioritizing to be done and software projects to manage and scope and ship, but the fact that your overriding mandate is "make the system faster|better|resilient" and not "delight fickle humans" basically eliminates many dimensions of complexity.
i am GLAD that we in infra are starting to wake up to the fact that we should treat engineers like human beings, not like robots. i am GLAD we are starting to incorporate principles of design into our shit.
i am just grappling with realizing how far behind we are, there.
in infra it is always our goal to work ourselves out of a job, by eliminating our jobs. so farther up and up the stack we must go.
good ops engineers don't work for mom and pop startups much now, we are all gravitating to solve category problems. we ops the whole internet.
if you look at the scope and sweep of infrastructure projects over the last decade or two, it is astonishing -- and encouraging! -- how much less time we now spend on must-dos, and how much time has been freed up for wanna-do's.
our teams are smaller, yet do way more. 🌈
the future of engineering is inexorably trending towards a world where all engineers ship product.
you'll have product engineers, and infra engineers. most of the world's great infra engineers (formerly "ops") will be clustered at infra companies, solving category problems.
there will be some SRE specialists who love to sit at the nexus of infra and product, consulting and encouraging and advising, and that will be increasingly a human-focused role for senior engineers with a deep grounding in infrastructure + strong social skills. (ex-managers 🙃)
ops will be continually redefined up and up the stack, because ops skills will no longer be someone else's job. you write the code, you ship the code, you debug the code in prod. period.
infra engineers build the subsystems upon which you skate in blissful ignorance.
we have zero dedicated ops engineers at honeycomb. we have a couple of formerly-opsen -- but they ship product too, and help the rest of the team with infra from time to time. ☺️
if you had told me this two years ago, i might not have believed you. shit is changing fast.
now i'm just thinking about things that were common wisdom three years ago when we started, and how they've changed.
"you can't put software engineers on call, they'll quit!" => "OBVIOUSLY software engineers should be on call, it's just about implementation details"
"monitoring is a solved problem, there's nothing left to be done in the space" => "all the players in APM, logs, metrics categories are rushing to unify their product offerings around wide structured events as the single source of truth (like honeycomb)"
"observability has three pillars and is a synonym for telemetry" => "observability is a technical term for how well you are able to understand what is happening inside your complex distributed systems, just by observing from the outside"
i wonder what else will be different in another three years. 🤔 maybe i will have learned how to help with sales without fucking it up whenever i try.
(and maybe santa's elves will have blue skin)
in conclusion, from honeyslack today via @alainadev, this gif.
You can follow @mipsytipsy.
Tip: mention @threader_app on a Twitter thread with the keyword “compile” to get a link to it.
Enjoy Threader? Sign up.
Threader is an independent project created by only two developers. The site gets 500,000+ visits a month and our iOS Twitter client was featured as an App of the Day by Apple. Running this space is expensive and time consuming. If you find Threader useful, please consider supporting us to make it a sustainable project.