Check various metrics and checks for Sidekiq. The following scripts:
running on your application. Pass the
--url URL option with the url to your monitor stats. Both scripts also accept a
--auth USER:PASSWORD option in case the monitor stats url sits behind basic authorization.
uses the native sidekiq JSON dashboard under
<mount point>/dashboard/statsMore info
Uses the maximum latency to figure out when to raise warnings. The latency is the time a job has been waiting in the queue (without being processed). It's generally a good metric to know when there's trouble with Sidekiq. There are some options you can pass to this script:
--warn SECONDS: Warn after job has been SECONDS seconds in a queue. The default is 120.
--crit SECONDS: Critical after job has been SECONDS seconds in a queue. The default is 300.
Generates total concurrency (
concurrency), total busy (
busy) and maximum latency (
latency) metrics for your sidekiq. A graph of
busy vs. concurrency is specially useful for capacity planning. The latency metric should let you know if something is wrong (or there
are spikes that don't trigger a warning or a critical). Accepts an
--scheme SCHEME option to know what scheme to append to the
metrics (the default is
Checks the dead queue and raises a CRITICAL if it's not empty. Jobs in the dead queue have failed over 25 times. Some organisations might consider this an issue to address and could benefit from this check.
Thanks to xxxxxx for contributing.