cloudsql_monitoring

Alerts for CloudSQL instances. These should probably be in the idf_deploy repo because we’re already declaring the instances and parameters there. But we decided to split our monitoring into a separate repo, so we re-declare them all here, because it seems easier to do that than to replicate the alerting channel and routing logic in the idf_deploy repo.

Available metrics are in the Google Cloud CloudSQL docs.

Inputs

Name	Description	Type	Default	Required
connections_threshold_percent	Notify when the percentage of existing connections to max_connections is greater than this.	number	70	no
cpu_threshold_percent	Notify when CPU usage is above this percentage, from 0 to 100.	number	80	no
disk_threshold_percent	Notify when disk usage is above this percentage, from 0 to 100.	number	80	no
google_project	Google Cloud project id. We need this because it is used in the database instance id that we need to query metrics for.	string	n/a	yes
instance_name	The name of the CloudSQL database instance to monitor.	string	n/a	yes
max_connections	The value of the max_connections Postgres parameter. This may be set explicitly, or it may be defaulted by Google based on the amount of memory of the instance. The defaults are listed in the Google Cloud docs (search the page for “max_connections”). The only way to know for sure is to query the db with SHOW max_connections;	number	n/a	yes
memory_threshold_percent	Notify when memory usage is above this percentage, from 0 to 100.	number	70	no
notification_channel_id	The ID of a Slack Google Cloud Monitoring notification channel to send notifications to.	string	n/a	yes