CPU Time - insight on what's taking time and how long should requests take generally?
I'm looking at a few of our apps that run on workers. I see various CPU Time metrics. I'm curious on the best way to dive deeper into this to see what requests are taking longer and figure out why. I'm also curious how long requests generally should take.
Is there anywhere I can read more on this? We're using the new "Standard" pricing model.
5 Replies
I thought the old limit was 50ms. Some of our requests are going up to 100ms. Many are close to 50ms. Is it calculated differently now? Can I dig into the data more somehow? Which requests are taking what amount of time for example.
The limit was 50ms now its 30000ms. You can however add a limit if you wanted to on a per worker basis.
And the only way to dig into the specifics of requests for how long CPU time wise would be if you have an enterprise account as only enterprise accounts have access to the logpush datasets that expose CPU time.
Thank you. One more quick clarification if you could... is it calculated differently than it used to? For example is my code actually taking this long to run or is it also taking into account waiting on fetching data or something? ~100 ms to essentially just fetch data from a CMS and render it seems wild
Its not calculated differently its the same.
And larger responses = more CPU time to parse the data so that could be the source of the higher CPU time requests.
Thank you very much! I'll dig in with what I can and figure it out