feat(symbolicator): Automatically add and remove projects to the LPQ based on their historical perf #28714
Conversation
| added = cluster.sadd(name=LPQ_MEMBERS_KEY, values=[project_id]) | ||
|
|
||
| if added: | ||
| # cry at sentry |
There was a problem hiding this comment.
don't think we need to do anything here? sentry will read the redis key and that's that
There was a problem hiding this comment.
iirc we do want to tell sentry something when projects are added and removed from the LPQ list, as a starting step to tracking the general behaviour of these tasks and when/which projects are matching the LPQ criteria. i assume that means sending something over via capture_message(), but could i be misinterpreting one of the goals or tasks related to that?
There was a problem hiding this comment.
oh, that way. yes we need this probably both as datadog metrics as well as some capture_message() probably. i misunderstood what you meant, the danger of working both on the product and using the product at the same time... 😄
16800e1 to
3b1ac0e
Compare
fb904e5 to
52936df
Compare
|
I've pulled out all of the code related to manipulating the list of projects into #28757 and rebased this PR on top of that one. |
f127836 to
c5eb2c9
Compare
823072b to
e365ab0
Compare
fc3e29e to
1644da0
Compare
This is a bit of a bloated PR, but this fills in basically everything else left besides the logic to determine whether a project belongs in the LPQ or not.
This does not include any logging to datadog; If that is needed and can be done, a follow-up PR can be opened to add that in.
More than half of the changes in this are simply tests, which will be broken up into classes for readability and modularity in a future PR. A significant number of lines are dedicated to simply declaring the API on a base class.