Client side metric queries

Q1. Following are our interpretetions of below metrics: workflow_task_execution_latency - latency of a workflow task which is shown in history of execution of workflow
workflow_task_execution_total_latency - latency for complete execution for 1 run id
workflow_endtoend_latency - all runs execution along with retries for 1 workflow id
Is this interpretation correct?
Q2. Are there any metrics which give number of retries of a workflow execution?
Q3. Any metric available to indicate number of workers on client side?
Q4. Is temporal web ui configurable for timestamp to be shown along with ms?

  1. workflow_task_execution_latency: execution time (latency) of a single workflow task (unit of execution of your workflow code), does not include workflow task retries
    workflow_task_execution_total_latency: total execution time (latency) of a workflow task, including things like retries

  2. you could use SDK workflow_continue_as_new metric i believe as workflow retries call ContinueAsNew internally

  3. No, you could use DescribeTaskQueue api to get the pollers for a specific task queue

  4. Not in v1. Our team is working on web ui v2 where this is addressed, you can try it out using this docker compose until its fully released.

1 Like

And what would be the interpretation for workflow_endtoend_latency?

workflow_endtoend_latency measures your workflow execution time from start to close.

1 Like