←back to thread

210 points dakshgupta | 1 comments | | HN request time: 0s | source
Show context
joshhart[dead post] ◴[] No.41842050[source]
[flagged]
candiddevmike ◴[] No.41842268[source]
Or the idea of an "interrupt handler". OP may find other SRE concepts insightful, like error budgets.
replies(1): >>41845477 #
wombatpm ◴[] No.41845477[source]
Error budget or recovery cost tracking goes a long way towards defeating the We never have time or money to do it right, but we’ll find time and money to fix it later mindset.
replies(1): >>41845618 #
dakshgupta ◴[] No.41845618[source]
I’m generally a strong believer in “if it’s not measured it’s not managed” so this seems like it would be useful to explore. I suspect it’s tricky to assign a cost to a bug though.
replies(1): >>41875840 #
1. wombatpm ◴[] No.41875840[source]
It’s tricky to assign impact, lost sales, revenue, customer satisfaction. Cost is just development and testing time.