Populate processInstanceId for execution jobs when invocationsPerBatchJob=1

[yanavasileva]

Acceptance Criteria (Required on creation)

• Populate process instance id (and process definition id where possible) for jobs
  • In case invocationsPerBatchJob=1 (jobConfiguration.getIds() != null && jobConfiguration.getIds().size() == 1)
    • Set the processInstanceId to the job.
    • Set the processDefinitionId to the job Edit: if the information is already available in the batch's configuration, at the moment only for restart batches. (That way the job logs will be visible in the Historic Process Definition page, Job log view.)
    • Code: org.camunda.bpm.engine.impl.batch.AbstractBatchJobHandler.createJobEntities(BatchEntity, T, String, List<String>, int)
• Clean up the process instance population that is already done for some type of batches in org.camunda.bpm.engine.impl.batch.AbstractBatchJobHandler.postProcessJob(T, JobEntity, T)

Hints

• Keep in mind that this implementation needs to be adopted by the users implementing custom batches

Links

• depended on
  • Track affected process instances by batch (when invocationsPerBatchJob=1) #4198
  • Expose filter option for instances that are not being part of batch when creating a modification #4200
  • Track process instances started by a restart batch operation #4201
  • Introduce notAffectedByModificationBatch filter criteria for process instances #4209
  • In Cockpit, add processInstanceId column to the Failed jobs and Job log view in Batches page #4216
  • (Optional) Introduce affectedByBatch filter criteria for historic process instance query #4215

Breakdown

Pull Requests

Beta Give feedback

1. feat(engine): Populating process instance id in jobs #4334
   +0
   
2. 4205 batch execution job changes process instance ID #4577
   ci:all-db ci:default-build +2
   

[yanavasileva]

Update on population of processDefinitionId for the batch execution jobs. The information is not available out of the box for most of the batches. With consideration of the performance (querying for each PI to fetch the process definition ID), we decided to populate the id only if it's already available (restart batch). We can iterate later on that if we receive user's feedback that this is needed.