Check various metrics and checks for Sidekiq. The following scripts:
- bin/check-sidekiq.rb
- bin/sidekiq-metrics.rb
require sidekiq-monitor-stats
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.
- bin/check-sidekiq-dead-queue.rb
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 sidekiq).
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.
- bin/check-sidekiq.rb
- bin/sidekiq-metrics.rb
- bin/check-sidekiq-dead-queue.rb
Thanks to xxxxxx for contributing.