OpsGenie Best Practice on scheduled alerts

johnathan_blanco

We have a lot of events that are scheduled through the day and we want to utilize Opsgenie to notify us if these events fail. 

IE at 5pm we get an email with an attachment that gets dropped to a network store. 
at 9pm 4 files go out via ftp and we get confirmation of success. 
at 4am we receive a set of files that we load on process. 
etc...

So essentially what I would want are scheduled 'soft' alerts that are ack/closed by the confirmations of success from each operation and if they don't get that confirmation by x-time to then escalate and notify the team. 

I didn't see a scheduled way to generate these event/alerts so I'm wondering what routes others have tried for this and what would be considered the best practical way of doing this. 

Thanks!

Tom Russell

@johnathan_blanco this isn't something Opsgenie can natively do without some external scripting/automation. There are several ways to do this, but each method will require some external system either sending an alert, an alert-close, or a heartbeat. Here are some suggestions:

1. Simplest would be to run a script at the checkpoints you mentioned above and create an alert if the condition is not met. We use our batch scheduling system to do this all the time.
2. If you want the visibility that you're waiting for completion, send an alert at start-of-day and suppress until the checkpoint time. As soon as the action is complete, have it send an alert close (use the alias field to match to the correct alert without having to save alert ID somewhere). This leaves the alert visible until it is verified and closed.
3. The Opsgenie Heartbeat is meant for situations like this. You can just set up a daily heartbeat and have your jobs send a heartbeat ping message when they complete. This may not work as well if there's a lot of variability in completion times (if your transfer finishes early one day, a 24hr heartbeat will expect it to finish at the same time the next day).

Just some thoughts...