bad API error message when using a not existing backend

[bkamin29]

When you use slo-generator in API mode, if you try to compute an SLI with a unknown backend (i mean not specified from the shared configuration), error message from the API is not clear :

Call :

curl -X POST -H "Content-Type: text/x-yaml" --data-binary @sli_get-organization_availability.yaml -H "Authorization: Bearer $(gcloud auth print-identity-token)" https://slo-generator-xxxxxxxx-ew.a.run.app

Answer:

Service Unavailable%         

Log seen from SLO Generator instance:

ERROR - Backend "cloud_monitoring/xxxxxxx" not found in config. Exiting.

maybe the error is not well catched ?

[lvaylet]

Hi @bkamin29,

Thanks for reporting this.

I think the error is caught properly at the SLO Generator level (hence the meaningful error message) but might not be propagated properly to Cloud Run. Unless the Cloud Run error has a different cause? Let me try to reproduce this and get back to you.

[bkamin29]

Hi @bkamin29,

Thanks for reporting this.

I think the error is caught properly at the SLO Generator level (hence the meaningful error message) but might not be propagated properly to Cloud Run. Unless the Cloud Run error has a different cause? Let me try to reproduce this and get back to you.

aaah, why not, in this case we can reproduce it by using the python module from a laptop ?