-
Notifications
You must be signed in to change notification settings - Fork 111
Open
Description
Version 0.18.0
When specifying multiple projects
/bin/stackdriver_exporter \
--google.project-id=gitlab-runway-staging,gitlab-runway-topo-svc-stg \
--monitoring.metrics-prefixes=spanner.googleapis.com \
--web.listen-address=:9256 \
--web.telemetry-path=/metrics
* [from Gatherer #2] collected metric "stackdriver_spanner_instance_spanner_googleapis_com_client_gfe_connectivity_error_count" { label:{name:"client_name" value:"spanner-go/1.83.0"} label:{name:"database" value
:"topology-service"} label:{name:"directpath_enabled" value:"false"} label:{name:"directpath_used" value:"false"} label:{name:"instance_config" value:"unknown"} label:{name:"instance_id" value:"topology-ser
vice-dev"} label:{name:"location" value:"asia-southeast1"} label:{name:"method" value:"Spanner.ExecuteSql"} label:{name:"project_id" value:"gitlab-runway-topo-svc-stg"} label:{name:"status" value:"Unavailab
le"} label:{name:"unit" value:"1"} gauge:{value:0} timestamp_ms:1763391420000} was collected before with the same name and label values
All errors are from Gather #2 so it looks like there may be a race condition when you don't specify individual metric prefixes.
I can work around this with:
/bin/stackdriver_exporter \
--google.project-id=gitlab-runway-staging,gitlab-runway-topo-svc-stg \
--monitoring.metrics-prefixes=spanner.googleapis.com/instance/cpu \
--monitoring.metrics-prefixes=spanner.googleapis.com/instance/storage \
--monitoring.metrics-prefixes=spanner.googleapis.com/instance/sessions \
--monitoring.metrics-prefixes=spanner.googleapis.com/api/request \
--monitoring.metrics-prefixes=spanner.googleapis.com/lock_stat \
--web.listen-address=:9256 \
--web.telemetry-path=/metrics
And then I don't see the error.
This issue only manifests with multiple projects using the --google.project-id option
Metadata
Metadata
Assignees
Labels
No labels