Checkpoint Notifications

[Matthew Bock]

We have a service that sends out an email at different stages of a large batching process. Certain areas are more critical compared to others and we have checkpoints built-in.

The scenario goes as follows: 

Batch Start: 17:00 - "Batch Started"

Checkpoint #1: 18:00 - "Step 1, Complete"

Checkpoint #2: 19:15 - 20:00 - "Step 2, Complete" 

Checkpoint #3: 21:00 - 22:00 - "Step 3, Complete"

Batch Finish: ??? - "Batch Completed Successfull"

What I would like to do is remove the 'email' to our team's distribution list entirely, and have OpsGenie handle these emails to notify as needed. 

My vision would be the "Start" notification is sent to OpsGenie at 17:00, since it's a start, no notification is sent, however, if no "Start" email is received by OpsGenie by 17:15, it needs to notify someone there is an issue. This would be the same process for the finish emails, they typically wrap up around 22:00, we would need the complete email to occur anytime after 22:00. 

Within each checkpoint, we do have subtasks that, when they fail DO go into OpsGenie and page out, as necessary. 

The checkpoint emails are in a similar boat as above, however, can drift 45 minutes depending on the delta of our workloads. They will never be ahead of their respective start time but their finish can vary. 

I'm not seeing an easy way to accomplish this task, there is no native integration with OpsGenie and the vendor in question, currently, this request is in their backlog to have wait times implemented, and if a process isn't completed within X minutes to notify. 

I've looked at the heartbeat feature, I don't see a way to configure the feature to only be active between time X and Y, for each checkpoint. 

[Shivam Naik]

Hi @Matthew Bock ,

Happy to help!

I would recommend submitting this to Support so that we can properly test out the proposed structure in conjunction with your current site set up. But while there isn't a timed reception filter (i.e. notifying the team if an alert isn't received by a certain time), you might able achieve this by filtering for time stamps using Action Filters, Notification Policies, and De-duplication wherein you could you have that Start alert arrive at a certain time, and not notify anyone via the aforementioned policies, and then using the Batch number as an Alias, you could possibly de-duplicate alert updates and actions throughout its lifespan.

Let us know if you have follow up questions!