Closed codefromthecrypt closed 9 months ago
ps we once had a test account and maybe I accidentally stopped it from running when updating CI config and/or removing travis.. I opened https://github.com/openzipkin/zipkin-gcp/issues/218 to get automated integration test with a real account working again, which would have prevented not noticing this. I ad-hoc tested the server once in the last couple months, but broke it after!
I found when the docker CI workflow was accidentally deleted.. several years ago. running the tests again in case the credentials still work.
ok rolling back CI as the google-cloud account was revoked, which is not terribly surprising the years
{
"status" : "DOWN",
"zipkin" : {
"status" : "DOWN",
"details" : {
"StackdriverStorage{zipkin-gcp-ci}" : {
"status" : "DOWN",
"details" : {
"error" : "com.google.auth.oauth2.GoogleAuthException: Error getting access token for service account: 400 Bad Request\nPOST [https://oauth2.googleapis.com/token\n{\](https://oauth2.googleapis.com/token/n%7B/)"error\":\"invalid_grant\",\"error_description\":\"Invalid grant: account not found\"}, iss: zipkin-gcp-ci-github@zipkin-gcp-ci.iam.gserviceaccount.com"
}
}
}
}
}
we'll just do ad-hoc integration tests until someone at google gives us an account again.
thanks for the look @reta!
I just tested the 2.2.2 tag, and it works like a charm. Thank you for the fix!
@provegard mentioned on gitter that we are getting an openssl error. I think the check is no longer valid after we switched to latest JRE + spring boot 3, as when I remove the check, the docker image (alpine JRE 21) works fine. So, I'm removing the check and in worst case, if someone has a misconfiguration, they get a late error in /health, but we don't get a false report for everybody else.
Traces created with https://github.com/openzipkin/brave-example/tree/master/armeria and self-tracing too. I ran the locally build docker image like so, configured with
Trace Agent
role