Re-routing escalationstions



Is there a way to re-route or customize the escalation path when the on-call support member rotation policy is not overridden due to some emergency.

Thanks in Advance,
Anand Kulkarni


Hi Anand,

Can you clarify your question? I’m not sure what you mean by “rotation policy is not overridden due to some emergency”. Are you trying to execute a different escalation based on existing or missing overrides in a schedule? If yes - why?


Hi Daniel,

Recently I faced an issue where OpsGenie failed to alert me on an notification. Here I got to know about the alert/issue from my support team. What if this repeats again and the escalation path is followed by OpsGenie directly without me getting notified on the alert? I wanted to counter this situation.
Well to know more about why OpsGenie did not notify me, I did consult one of the support members from OpsGenie and they could not find the log files for the same as I highlighted this issue more than a week later from its occurance. I was told that only 10 days of logs will be available.


Hi Anand,

The concept of escalation exists due to missed notifications (for any reasons). If the alert is a low priority one and it shouldn’t be escalated, you should be executing a different notification flow for those alerts. There is no way to stop an escalation because the alert has been missed, as that is the core reason why they exist.

However, the system shouldn’t fail to notify you. A single notification channel can fail, but a combination of those should lower the risk significantly.

Here is what I would do:

1.) Double-check my personal notification rules and verify, if I’m using all 4 channels in a timely manner (email, push, sms, voice)

2.) Setup a “Repeat” mechanism, which will restart the whole flow if the alert is still unacknowledged at the end of the notification chain.

3.) Consider increasing the time between the escalation steps, of course depending on the priority of the alert.

+1: Share the alert ID with the support team in time if anything strange happens, so they are able to verify the current alerting logic and check the logs which are available for 13 days. :slight_smile:

Does this help?