Open sdcote opened 8 years ago
This can be resolved by writing Batch Manager nuggets to query the status of jobs.
We now have a statistics board and can generate events in the event queue and have the loader update states for each of the components it loads.
The DefaultLoader
should probably do a little more in the watchdog thread, maybe placing some indication of a flakey component in the stats board.
This is more of an enhancement; the needs to be a uniform way to determine the status of a component.
Case; a scheduled job is failing to complete but the only way to know this is to look in the logs.
Case: a managed component failed to load and simply exits. The watchdog thread restarts the component only for it to continually fail.
Case: The management server should have a dashboard for component health and metrics.