You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In version 1.7.0 metric elasticsearch_cluster_health_up was removed as "not useful" (per commit acddc580d6b807d3400a4d711cf2516049912f39). We have been using this metric as generic "exporter is up and able to collect (cluster health) data from Elasticsearch" metric to alert us when the exporter is unable to do so, using:
min without (endpoint, instance, service) (elasticsearch_cluster_health_up) == 0 or absent(elasticsearch_cluster_health_up)
We could use the up metric and check the absence of some other metric for this alert rule, but this provides no indication what kind of response the exporter receives from Elasticsearch. Can you elaborate on why elasticsearch_cluster_health_up was deemed not useful and suggest an alternative, or can the metric be reinstated?
The text was updated successfully, but these errors were encountered:
In version 1.7.0 metric
elasticsearch_cluster_health_up
was removed as "not useful" (per commitacddc580d6b807d3400a4d711cf2516049912f39
). We have been using this metric as generic "exporter is up and able to collect (cluster health) data from Elasticsearch" metric to alert us when the exporter is unable to do so, using:We could use the
up
metric and check the absence of some other metric for this alert rule, but this provides no indication what kind of response the exporter receives from Elasticsearch. Can you elaborate on whyelasticsearch_cluster_health_up
was deemed not useful and suggest an alternative, or can the metric be reinstated?The text was updated successfully, but these errors were encountered: