[Stack Monitoring] Document timeout settings #34255
Labels
Feature:Stack Monitoring
Module:beat
The beat Metricbeat module
Module:elasticsearch
Elasticsearch Beats modules
Module:kibana
Kibana Beats modules
Module:logstash
Logstash Beats modules
Stalled
Team:Monitoring
Stack Monitoring team
Within Metricbeat there is a base timeout of 10 seconds which correlates to the default
period
setting, this is to avoid calls taking longer than the period to ensure the data is as fresh as the configuration expects.There are cases where we need to extend the
period
, such as when an endpoint takes longer to reply for "good" reasons. One example of this is the_cluster/stats
endpoint of Elasticsearch for a large cluster state. But, simply setting theperiod
to 20 seconds (for example) isn't enough, since we'll hit the request timeout limit instead.The below settings can be used to raise the timeout:
These settings are not well documented (and when to use them is also not documented).
AC
The text was updated successfully, but these errors were encountered: