Do all the things like ++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatarSign Up
Sounds like CYA should be job one.
Establish a baseline latency below which you have no control and bring it up every time someone comes at you with a number you aren't responsible for.
Outside of that, go out of your way to communicate back-end failure in the ui:
- toasts that indicate an operation is taking longer than expected
- fast get timeouts with connection issue notifications to the user
- skeleton placeholders
- slow network indicators
- allow the users a means of sending status telemetry per issue containing diagnostic info that are autorouted to the backend devs so you don't end up triaging
- use tools like logrocket to see what users are doing in real time per session and get error and transaction diagnostic interception
Lastly, keep p-value statistics for call lengths, and p-values for reverse proxy/post Fe query transit to tell the whole story.
tirthaguha15Performance Optimization: when you spend 3 days, to improve site loading by 0.5 secs.
Chippy2"I'm not ugly. I just reduce my graphics to have better performance."
dfoianu2Pessimist: a O(2^n) algorithm's performance decreases exponentially as input increases. Optimist: a O(2^n) al...