chore: add profiling labels for pprof analysis #19232
Conversation
PProf labels segment the code into groups for determing the source of cpu/memory profiles. Since the web server and background jobs share a lot of the same code (eg wsbuilder), it helps to know if the load is user induced, or background job based.
|
I'm sure the |
enterprise/coderd/coderd.go
Outdated
| @@ -903,7 +903,8 @@ func (api *API) updateEntitlements(ctx context.Context) error { | |||
| } | |||
|
|
|||
| api.AGPL.PrebuildsReconciler.Store(&reconciler) | |||
| go reconciler.Run(context.Background()) | |||
| // TODO: Should this context be the app context? | |||
There was a problem hiding this comment.
What is the "app" referring to here?
There was a problem hiding this comment.
api.ctx.
I just noticed the prebuild routine does not get the app context shutdown.
| } | ||
|
|
||
| const ( | ||
| ServiceTerraformProvisioner = "terraform-provisioner" |
There was a problem hiding this comment.
Follow-up: would be cool to label pubsub as well.
There was a problem hiding this comment.
Which pubsub? The database pubsub listeners?
coder/coderd/database/pubsub/pubsub.go
Line 246 in cb1ec21
| @@ -333,6 +333,7 @@ func New(ctx context.Context, opts *Options) (*Server, error) { | |||
| r.Use( | |||
| // TODO: @emyrk Should we standardize these in some other package? | |||
| httpmw.Recover(s.Logger), | |||
| httpmw.WithProfilingLabels, | |||
There was a problem hiding this comment.
Worth discriminating on ws vs https traffic?
There was a problem hiding this comment.
Oh that is a really good idea. They should have that upgrade header iirc. Will add
There was a problem hiding this comment.
request_type: Total 730.0ms
550.0ms (75.34%): http
180.0ms (24.66%): websocket
service: Total 810.0ms
730.0ms (90.12%): http-api
80.0ms ( 9.88%): terraform-provisioner
PProf labels segment the code into groups for determing the source of cpu/memory profiles. Since the web server and background jobs share a lot of the same code (eg wsbuilder), it helps to know if the load is user induced, or background job based.
Example output: