I'm thinking about integrating Jira with our Datadog setup to automatically create tickets for on-call issues, with relevant information injected into them. We're specifically looking into this for APM as well as logs-based and security monitors. One big concern is the issue of monitor flapping—how do we prevent Datadog from creating hundreds of tickets over a weekend that would need to be manually cleaned up later? Is there a way to configure the integration to check for existing Jira tickets with the same subject or summary? Also, I'd love to hear about any personal experiences you've had with Datadog and Jira integration, the good and the bad. I've read the documentation, but real-world insights would be invaluable!
3 Answers
The flapping problem is definitely something to watch out for. We moved away from a direct Datadog-to-Jira integration and started routing alerts through an incident management layer instead. This setup manages deduplication and correlation, only cutting tickets when there's a genuine need for human attention. It really cuts down on noise compared to letting Datadog just fire off tickets to Jira directly.
Absolutely! We’ve been using the Datadog-to-Jira integration in production for a few years now. It definitely works, but you need to set some ground rules to avoid too much noise. Instead of relying on Jira for flapping issues, fix them within Datadog by increasing evaluation windows and using composite monitors. It's better to create one issue per alert and update it rather than constantly opening new tickets. To manage duplicate concerns, we standardized Jira summaries to include service and monitor IDs. This way, Datadog can update existing issues instead of creating new ones unless absolutely necessary. Ideally, you want to limit this integration to just your critical alerts.
Before you dive into the integration, make sure your monitors are in good shape. If they flap, Datadog will create a ticket for every single state change. Opt to set your notification conditions to just alert, and consider adding an evaluation buffer. Just a heads up, Jira doesn't have native deduplication, so you'll need to handle that on your end. A lot of teams also bring up using Domo for tracking ticket trends and alert volumes over time without too much hassle.
Related Questions
Can't Load PhpMyadmin On After Server Update
Redirect www to non-www in Apache Conf
How To Check If Your SSL Cert Is SHA 1
Windows TrackPad Gestures