Alert Notification to track UNKNOWN status in a HA failover
Alerts are generated when the status of services/components changes from OK to UNKNOWN
And alert notification configured to send email for a specific alert group and for a specific severity.
In case of a YARN HA, when active resource manager fails over, ambari generates alert for this resource manager. The alert status changes from OK to UNKNOWN only for the failed resource manager while the active Resource Manager continues to show status OK. Alert notification can be configured to track the UNKNOWN status and send email. Following is an example for YARN HA but same would apply for Namenode HA:
1. Create a custom alert group to track Resource Manager alerts.
a. From alerts menu, click on Actions->Manager Alert Groups
b. Create new alert notification. Choose severity as UNKNOWN and group as group created in step 1.
c. Click on save.
3. Email based on default template will be sent whenever the status changes to UNKNOWN to the email address setup in step 2.