JEF Executor metrics
JEF outputs metrics for each of the JEF services as well as each of the Action Executors services (e.g. Job Async Executor, Transcode Resource etc.). A graph dashboard is available to represent collected metrics.
If you develop a new ActionExecutor microservice, it will be automatically added to Dalet Flex metrics and the JEF dashboard.
Please contact your Dalet representative for more information.
JEF Executor properties (highlights)
|job.resume||TRUE||If this property is enabled for non-ActionProgress supported jobs in a TIMED_OUT state, the watcher will try to call `resume()` after the 2nd deadline expires (job.watcher.scheduler.interval). <br><br> If this property is enabled for those jobs in a QUEUED state, the watcher will try to move them to RUNNING state after 2nd deadline expires (job.watcher.scheduler.interval). <br><br> If this property is enabled for jobs in a QUEUED state, the service will try to move them to a RUNNING state when the microservice is started.|
|job.watcher.scheduler.interval||600000||Watcher cadence. This is how often the watcher runs.|
|job.queued.lastUpdated.threshold||300000||This property is used to find out potential QUEUED jobs that have stalled. Jobs in a QUEUED state that are not moved to a RUNNING state will be candidates for watcher processing.|
|job.timedOut.lastUpdated.threshold||600000||This property is used to identify potential jobs in a RUNNING or TIMED_OUT state that have become stalled. Jobs that are in a RUNNING state and are failing to progress, may be moved into a TIMED_OUT state by the watcher. Jobs that are in a TIMED_OUT state and are failing to update, may be moved into a FAILED state by the watcher. Alternatively the call to `resume()` will be made.|
|executionStep.numberOfExecutors||30||The number of threads available in each iteration to process actionProgress.</td>|
|executionStep.max.jobs.limit||5000||This property is used to specify the maximum number of actionProgress running in a single job executor microservice (actual total number may vary during the execution and may always be accurate in relation to this threshold).|
|job.numberofexecutors||40||The number of threads available to process job requests coming from the message queue (JobRequest).|