Aller au contenu

Overview#

Metrics#

List of metrics retrieved#

Currently we are retrieving metrics from FE components:

  • RDS

Metrics are prefixed with flexible_engine_.

Example with RDS component : flexibleengine_sys_rds_active_connections.

Note

For more information: Summary of available metrics and Details of available metrics

Add a new component to monitor#

envs-ng#

In cloudeye-exporter.cue modify scrapeServices tab and add component to monitor.

Deploying#

Deploy as usual with Trackbone.

Add rules#

We can deploy several types of rules :

  • to alert on the behavior of the component

  • to alert when component metrics are missing : this can happen for example when the servicemonitor is not there.

There is already a file in repo applications/caascad-prometheusrules : cloudeye-exporter.cue.

Example alert for missing metrics for RDS component : 

absent(flexibleengine_sys_rds_active_connections{job=\"cloudeye-exporter\", namespace=\"cloudeye-exporter\", service=\"cloudeye-exporter\", cc_prom_source=\"\(zone)\"}) and on(cc_prom_source) caascad_cluster_info{cloud_provider=\"fe\"} == 1

Note

The second part of the expression is used to trigger the alert for FE clusters.

Architecture#

Cloudeye-exporter is deployed in cloudeye-exporter namespace.

We deploy two servicemonitors :

  • one to retrieve the metrics of the application itself : cloudeye-exporter-metrics

  • another to retrieve the metrics of the FE components : cloudeye-exporter