Please help me analyze the possible causes:
-
When the platform calls
runtimeService.createMessageCorrelation(messageName)
.processInstanceBusinessKey(BpmnModelParser.getProcessActId(workflowId))
.setVariableLocal(WorkflowConst.TASK_STATUS_TYPE, taskStatusType)
.correlate()in a scheduled task, it gets stuck for 1.5 hours and then automatically recovers; -
This code is called within a for loop;
-
Investigation revealed that the
ACT_RU_EVENT_SUBSCRtable has accumulated a large amount of historical data; -
The current Camunda version is 7.19.0;
-
Database investigation showed long-running transactions at the time, but no long-term database locks;
-
Other platform functions are normal; only this scheduled task is affected, and other scheduled tasks also operate normally;
-
This phenomenon has occurred multiple times with varying recovery times observed: 1.5h, 1h, 3h, 5h.