PEGA0131 alert: Detected slow data flow component
The data flow engine detected a slow component in a data flow run and has been unable to process the current record for at least 5 minutes. The data flow engine will continue to process the record as an interruption in processing might have undesired side effects, such as:
- Connections to the database or other resources might not be closed.
- Currently held synchronization locks might not be properly released.
Reasons for the alert
The component that is involved in the processing of the current record is slow or blocked.
You might see one of the following messages, depending on whether Pega Platform™ was able to identify the component that caused the problem:
When the slow or blocked component is known:
Slow component ['<ComponentName'] with id ['<ComponentId>'] detected in data flow run ['<RunId>'] on data flow ['<ComponentDataFlow.AppliesToClass>'.'<ComponentDataFlow.RuleName>'] and is still running on thread ['<SlowThreadName>']
When the slow or blocked component is unknown:
Slow component detected in data flow run ['<RunId>'] and is still running on thread ['<SlowThreadName>']
Investigate the stack trace that is attached to the alert. From the stack trace, you can determine whether the component is slow or blocked due to one of the following reasons:
- Infrastructure issue - Input/output communication with a database or other resources might be slow or blocked due to a large number of requests or resource unavailability. The stack trace often contains the name of the resource used. Resolve the infrastructure issue to prevent slow or blocked data flow runs.
- Application design issue - An activity, function, or data flow that you trigger from a data flow might contain a component that is slow or blocked. For example, a step in an activity to send an email might be slow or not responsive due to the overloading of the email server. Redesign the application or add extra error handling for such cases to prevent slow or blocked data flow runs.
If the data flow engine cannot process the next record in more than 15 minutes, the health check marks the node as unhealthy.
- In environments with Pega health check enabled, such as Pega Cloud, a graceful shutdown is initiated.
- In on-premises environments, if the data flow engine still cannot process the record after you fix the slow component, restart the node.