Need help in understanding poller/worker start metrics

Community Support
,
1

Hi,

We have a K8S deployment where we have multiple pods of our application that is hosting activity/workflow workers on different task queues with maxConcurrentWorkflowTaskPollers and maxConcurrentActivityTaskPollers as 64.

  1. Let say, I have X no. of pods of my application, I am able to see X no. of pollers on a particular task queue from web ui, but temporal_poller_start_total metric is showing values as zero for each pod in prometheus/grafana. As per the documentation, this metric indicates the counter of worker entity pollers that started, but does not seem to reflect in the scraped metrics.
    Also, is there any other metric in Java SDK to check no. of pollers an a task queue?

  2. Even temporal_worker_start_total metric is showing values as zero for each pod in prometheus/grafana.

I am keen in understanding worker and poller related metrics and check those numbers for each of the pods of my application. Please explain.

2

Looks like I figured out the issue. Problem looks with prometheus scraping, because I am able to see metrics with appropriate values from the “actuator/prometheus” endpoint of my microservice. I will investigate the issue further with my prometheus setup.