Yes. There is a vast wilderness of problems and weirdness that are not (yet) serious enough or customer impacting enough to wake someone up.
Which is why you should spend some time looking at production Every. Day., and following your curiosity where it takes you.
Which is why you should spend some time looking at production Every. Day., and following your curiosity where it takes you.
It was established monitoring doctrine that "you shouldn't need to go look at graphs all the time, the system should alert you when it's broken."
No more.
(Relevant: charity.wtf/2019/09/20/lov…)
No more.
(Relevant: charity.wtf/2019/09/20/lov…)
You really do need to be in there looking at the code you are writing, through the lens of your instrumentation. Every day. Just poking around.
I don't care how good your test coverage is; if you aren't watching users use it in prod, you aren't done.
I don't care how good your test coverage is; if you aren't watching users use it in prod, you aren't done.
Think of it as "brushing your teeth" for systems. Find and scrub all the tiny little problems while they are tiny, before your users find them.
Compound this approach over time and that's how you get a clean, stable system that is well understood. 🌷
Compound this approach over time and that's how you get a clean, stable system that is well understood. 🌷
