cloudfoundry-community
Intermittent errors in the nozzle regarding project_id
After a successful apply changes to deploy the stackdriver nozzle, we were starting to lose some logs from the nozzle intermittently. After looking at the logs from the nozzle, we started to see errors of the form:
"error": "rpc error: code = InvalidArgument desc = One or more TimeSeries could not be written: Field timeSeries[0].resource.labels.project_id had an invalid value of \"our-pcf-project-1234\": if present, must be the project ID in the request name.:
I've been trying to dig around to see what might be the issue. The only strange thing we are doing is that we are using a service account in on project, to be able to post logs into a different central stackdriver project for all of our pcf foundations. So the project_id field that we input during the installation of the tile is different from the project_id that is provided within the service account that we are using.
Could this be causing these issues?
Hi @shinmyung0 , thanks for reporting this!
We've been researching intermittent log outages, so it might be a difference in project ID, or it might be another issue.
Currently I'm focusing on merging the develop branch as my top priority. Once that's done I'll circle back to the other issues.
In the meantime, @evandbrown might know more details.
Currently trying it with the same project ids for both the target project as well as the service account and seeing if that helps.