I started collecting cpu profiles from some services written in Typescript, and when I want to look at them in the UI, it intermittently fails by turning the whole page black, and the following javascript error:
app.e9df57e….js:2 TypeError: i[e.getLabel(...)].push is not a function
at app.e9df57e….js:69:66536
at Vj.initLevels (app.e9df57e….js:69:66730)
at Vj.getCollapsedMap (app.e9df57e….js:69:65990)
at app.e9df57e….js:47:866
at oc (app.e9df57e….js:2:1113477)
at Al (app.e9df57e….js:2:1133422)
at ll (app.e9df57e….js:2:1125815)
at Yo (app.e9df57e….js:2:1067057)
at app.e9df57e….js:2:1132254
at vl (app.e9df57e….js:2:1132259)
app.e9df57e….js:2 Uncaught TypeError: i[e.getLabel(...)].push is not a function
at app.e9df57e….js:69:66536
at Vj.initLevels (app.e9df57e….js:69:66730)
at Vj.getCollapsedMap (app.e9df57e….js:69:65990)
at app.e9df57e….js:47:866
at oc (app.e9df57e….js:2:1113477)
at Al (app.e9df57e….js:2:1133422)
at ll (app.e9df57e….js:2:1125815)
at Yo (app.e9df57e….js:2:1067057)
at app.e9df57e….js:2:1132254
at vl (app.e9df57e….js:2:1132259)
Sometimes I can see the profiles, then I change the time window, and then UI fails again.
To Reproduce
Steps to reproduce the behavior:
Start Pyroscope 1.9.1
Collect cpu profiles from a Typescript service using Alloy pull mode (@pyroscope/nodejs v0.4.1)
View cpu profiles for that service in the Pyroscope UI
Expected behavior
I would expect the UI not to fail
Environment
Infrastructure: kubernetes
Deployment tool: helm
Additional Context
I am using the alloy configuration provided by the helm chart (slightly modified for godeltaprof profiles). The relevant generated pyroscope.scrape config looks like this:
Describe the bug
I started collecting cpu profiles from some services written in Typescript, and when I want to look at them in the UI, it intermittently fails by turning the whole page black, and the following javascript error:
Sometimes I can see the profiles, then I change the time window, and then UI fails again.
To Reproduce
Steps to reproduce the behavior:
Expected behavior
I would expect the UI not to fail
Environment
Additional Context
I am using the alloy configuration provided by the helm chart (slightly modified for godeltaprof profiles). The relevant generated pyroscope.scrape config looks like this:
The logs of Alloy are clean, and so is the debug page for this component.
Logs from query-frontend:
Logs from query-scheduler:
Logs from querier:
Logs from ingester:
Logs from store-gateway: