I have used helm-chart for temporal deployment. I can access the frontend and able to run workflows, however, metric endpoint is not exposed or may be the metric process is not running on the custom port 9090.
Do you by chance disable prometheus when you do helm install? (–set prometheus.enabled=false)
@tihomir Thanks for the response. I am using different prometheus chart not the one used in temporal helm-charts.
I can see the service monitor deployed on k8 cluster but
I don’t see metric port exposed on the pod helm-charts/server-deployment.yaml at master · temporalio/helm-charts · GitHub
However, when I run temporal through docker-compose I can see the metric port 8000 exposed.
Prometheus will only scrape the the data if the port is exposed.
After following the thread Configuring external promethrus and grafana - #3 by tihomir
This issue is now resolved. The issue was because of upgrade from 1.12 to 1.17 and since server_configmap.yaml is different in both, hence the metric port was not exposed.
1 Like