Logging automatic retries / Actual Retry Count in the Cockpit?

Boris-S · October 21, 2022, 2:11pm

Hi,
we are looking for a way to see if a task had automatic retries.

Is there an option to see the actually done retries on a task in the cockpit?

Or can we turn on logging / is the number of the current retry count already logged somewhere? We can’t seem to find it in our quite extensive splunk logs…we are pretty sure we have retries but it would be great to have camunda tell us

thank you,
Boris

jonathan.lukas · October 21, 2022, 9:20pm

Hello @Boris-S ,

you could implement a custom FailedJobRetryCmd.

Am example can be found here:

github.com

camunda-consulting/camunda-7-code-examples/blob/master/snippets/offline-schedule/src/main/java/com/camunda/consulting/scheduleRetryOfFailedJobs/OfflineSchedulePlugin.java

package com.camunda.consulting.scheduleRetryOfFailedJobs;

import org.camunda.bpm.engine.impl.cfg.AbstractProcessEnginePlugin;
import org.camunda.bpm.engine.impl.cfg.ProcessEngineConfigurationImpl;

import com.camunda.consulting.OfflineSchedule;

public class OfflineSchedulePlugin extends AbstractProcessEnginePlugin {
  private final OfflineSchedule schedule;

  public OfflineSchedulePlugin(OfflineSchedule schedule) {
    this.schedule = schedule;
  }

  @Override
  public void preInit(ProcessEngineConfigurationImpl processEngineConfiguration) {
    processEngineConfiguration.setFailedJobCommandFactory(new ScheduledFailedJobCommandFactory(this.schedule));
  }
}

I hope this helps. And sorry I didn‘t have this in mind in the afternoon

Jonathan

fml2 · October 22, 2022, 10:27pm

If you use the enterprise version of Camunda, then you have the history view of a process instance. There you have a “Jobs Log” tab where you can see how and when jobs were executed and what the result was (success or failure).

Since this is displayed I assume that the data is present in the DB. So, if you happen to use the community version, you can query the data via API. I don’t know which service can deliver this data.

Boris-S · October 25, 2022, 3:54pm

thanks @jonathan.lukas @fml2

I now found a job which failed and got successfully retried (I guess"Due Date" is set after a failure for the next retry?)

However, I would have expected “Retries” to be decremented… but it stayed the same.

(A runtime Exception is thrown during the Delegate Call).

jonathan.lukas · October 25, 2022, 8:35pm

Hello @Boris-S ,

did you already apply custumization to the retry failed job command?

Jonathan

Boris-S · October 26, 2022, 5:56am

Good point, not that I am aware of anything - but we use some common framework/configurations, I need to check if something is done there…

Boris-S · October 26, 2022, 7:57am

@jonathan.lukas Are you suggesting that the behaviour I am observing is customized?

jonathan.lukas · October 27, 2022, 5:24am

Hello @Boris-S ,

when reproducing it, I can see that number should decrease for each retry.

Maybe the from me suggested defaultNumberOfRetries should be replaced with failedJobRetryTimeCycle and an ISO-8601 expression.

Jonathan

Boris-S · November 1, 2022, 3:20pm

Well, I don’t know, I haven’t implemented anything… I was assuming we have standard Camunda behaviour

thx,
Boris