When monitoring a process engine for bottlenecks, it is useful to look at the queue length of pipelines and the time it takes to process steps between pipelines. Processes in the pipeline queues refresh every 30 seconds.
Process engines use the following pipeline types:
Defines the number of active processes currently loaded on this process engine. This includes processes with a status of “Running”, “Error”, and “Aborting”, but not those with a status of “Completed” or “Aborted”.
Displayed only in the Internal Engine Queue section.
Defines the number of step instances that are currently waiting for events, such as a user action to submit a change in CA Clarity PPM. When events registered by some step instances on the Event Wait list come, these step instances are pushed to either the Pre-condition queue or Post-condition Transition queue.
Process engines register only for events they are interested in. For example, if the engine is processing a post-condition for an active project that needs to go to the next step, the engine will register for the Project Update event.
Displayed only in the Internal Engine Queue section.
Defines the number of step instances that are waiting to be retried.
When a process engine processes a step instance, the engine may run into database deadlocks. When database deadlocks occur, the engine will place these step instances on the Retry Wait List.
Displayed only in the Internal Engine Queue section.
Defines the queue length is the number of step instances in the state “Ready to Evaluate Pre-condition” that are waiting to be processed by a Pre-condition pipeline.
Defines the queue length is the number of step instances in the state “Ready to Execute Action” that are waiting to be processed by an Action Execution pipeline.
Defines the queue length is the number of step instances in the state “Ready to Execute Post-condition” that are waiting to be proceed by a Post-condition Transition pipeline.
To simply monitor or clear the pipelines, you can do the following:
From the Process Engines page, click a process engine name to view the engine's internal queues page. The Engine Internal Queues page gives status information about the internal queues and the number of pipelines (if applicable).
The following describes pipeline metrics:
Indicates whether the process engine is working.
Defines the percentage of processing load across all internal queues.
Defines the percentage of engine time in the last time window that was used for processing. A time window is a variable time slice driven by many factors, such as load on each engine.
Specifies the time when the engine started.
Specifies the total time the engine used for processing since the engine start time.
Defines the names of the queue types in the process management infrastructure. The only queue types that you can configure are the pre-condition, action, and post-condition queues.
Defines the number of requests that are currently pending for this queue type.
Defines the number of configured pipelines per queue type.
You can add more pipelines to reduce bottlenecks in pipeline queues, or remove pipelines from each queue. You can have up to five pipelines per queue.
To add more pipelines
The Process Engines page appears.
The Configure Pipelines page appears.
Additional pipelines are added to the queue.
| Copyright © 2010 CA. All rights reserved. | Email CA Technologies about this topic |