0.12.0: Optional error class label in failed jobs counter

Added

• Optional capture of error class for failed jobs counter.

  Set label_for_error_class_on_sidekiq_jobs_failed to true to add error label to sidekiq_jobs_failed_total metric.

  Pull request #34 by [@niborg]

Changed

• Stop including development-related files into packaged gem to avoid confusing users or software tools. [@Envek]

0.11.0: Retry set segmentation

Added

• retries_segmented_by_queue configuration setting to allow segmentation of retry count by queue.

  It is disabled by default as it requires to iterate over all jobs in the retry set and may be very slow if number of retries is huge.

  Pull request #32 by [@SxDx]

0.10.0: Rerouted jobs monitoring

Added

• New metric sidekiq_jobs_rerouted_total_count to measure jobs that on enqueue were pushed to different queue from the one specified in worker's sidekiq_options. See #30. [@LukinEgor]

Fixed

• In sidekiq_jobs_enqueued_total_count track real queue that job was pushed into, not the one specified in sidekiq_options (sometimes they may be different). See #30. [@LukinEgor]

0.9.0: Support Prometheus official client multi-process mode

It is needed when you can't scrape every Sidekiq process directly and want to aggregate metrics from all worker processes on a server and expose them from within a single dedicated process.

Added

• Configuration setting to declare worker in-process metrics outside workers.

  It can be needed for official Prometheus client in multi-process mode where separate process expose metrics from Sidekiq worker processes.

• most_recent aggregation for all cluster-wide gauges.

  It is also needed for official Prometheus client in multi-process mode to reduce number of time series.

0.8.2: Access and change gem config via `Yabeda::Sidekiq.config`

Added

• Ability to programmatically change gem settings by calling writer methods on Yabeda::Sidekiq.config. [@Envek]

  Usage is quite limited though as you need to do it before Yabeda.configure! is called.

0.8.1: Compatibility with Sidekiq 6.2.2+

Fixed

• Compatibility with Sidekiq 6.2.2+ due to renamings in Sidekiq's undocumented API that yabeda-sidekiq uses. See mperham/sidekiq#4971. [@Envek]

0.8.0: Currently running jobs max time metric and global metrics collection switch

Added

• sidekiq_running_job_runtime metric that tracks maximum runtime of currently running jobs. It may be useful for detection of hung jobs. See #17. [@dsalahutdinov], [@Envek]

• Setting collect_cluster_metrics allowing to force enable or disable collection of global (whole Sidekiq installaction-wide) metrics. See #20. [@mrexox]

  By default all sidekiq worker processes (servers) collects global metrics about whole Sidekiq installation.
  Client processes (everything else that is not Sidekiq worker) by default doesn't.

  With this config you can override this behavior:

  • force disable if you don't want multiple Sidekiq workers to report the same numbers (that causes excess load to both Redis and monitoring)
  • force enable if you want non-Sidekiq process to collect them (like dedicated metric exporter process)

0.7.0: Apply tags from yabeda_tags to all metrics inside a job

Changed

• Tags from yabeda_tags method are applied to all metrics collected inside a job, not only sidekiq-specific. See #14. @Envek

0.6.0: Custom runtime tags at job level

Added

• Ability to override or add tags for every job via yabeda_tags method. @Envek

New job latency histogram, queue latency gauge renamed

Added

• New sidekiq_job_latency histogram to track latency statistics of different job classes. #9 by @asusikov

Changed

• BREAKING CHANGE! Renamed sidekiq_jobs_latency gauge to sidekiq_queue_latency to better describe its purpose and differentiate with the new histogram. #9 by @asusikov