Provide logstash_stats.queue.events_count as a "queue depth" metric in the logstash monitoring UI.
Describe a specific use case for the feature:
Today we get "Events ingested" which combined with CPU can give a rough idea of where to look for overwhelmed logstash nodes.
But we've found logstash_stats.queue.events_count metric (already provided from logstash in the metrics index documents) to be a much better indicator of a node that needs to be removed from rotation to let drain.
We can build visualizations for now, but could be good to get this into the default monitoring view.
Describe the feature:
Provide
logstash_stats.queue.events_count
as a "queue depth" metric in the logstash monitoring UI.Describe a specific use case for the feature:
Today we get "Events ingested" which combined with CPU can give a rough idea of where to look for overwhelmed logstash nodes.
But we've found
logstash_stats.queue.events_count
metric (already provided from logstash in the metrics index documents) to be a much better indicator of a node that needs to be removed from rotation to let drain.We can build visualizations for now, but could be good to get this into the default monitoring view.
Related to https://github.com/elastic/kibana/issues/18407