Terraform for alert sources

You can now manage your alert sources and attributes using our Terraform provider 🎉

This is useful if you:

• Want to use version control to manage changes to your alert source configuration
• Have a set of similar alert sources you’d like to generate configuration using a template
• Want to use the alert source’s token in your other Terraform configuration
• Want to apply complex access controls (e.g. members of each team can only edit their own alert sources)

Alert routes can now escalate to specific users and use multiple expressions

We’ve added some powerful tooling to your alert routes on how you can escalate. Some example use cases include:

• Automatically escalating to a specific user who broke something based on an offending pull request.
• Escalating an alert route to both the Team defined on the alert, as well as the Feature on it. If both of those point to the same escalation path, we'll deduplicate and only notify once.

As part of this, we've also made some improvements to the escalation timeline. Because you can now dynamically escalate to any user, we need to handle that user potentially being deactivated. If this happens, we'll show you in the timeline that a user within an escalation wasn't contacted as they were deactivated.

🚀 What else we’ve shipped

💅 Alert priority in alert sources doesn’t need to be set anymore, so you fallback to your chosen global default without having to redefine a fallback in each source

💅 A live call's escalation timeline will now tell you if it failed due to the caller hanging up

💅 Allow editing of more automatically set timestamps from Slack

💅 In the alerts configuration view, we've made it clearer when a private alert route only creates escalations

🐛 Fixed an issue where the escalation timeline on web was grouping notifications for a single user irrespective of the time the notification happened

🐛 Fixed an issue where an escalation would continue for a user if it was acknowledged as one of many escalations and the only notifications for that user had been rate limited

🐛 Fixed an issue where alerts with no attributes were excluded from insights

🐛 Fixed an issue where all escalations on an alert's details page were showing the same user’s avatars

🐛 Fixed an issue on Android where user's couldn't dismiss DND settings due to being installed in a work profile

🐛 It's now a lot faster for larger organizations to load their escalation policies from PagerDuty in order to import them

🐛 Fixed a bug where the date heading was in the wrong timezone in escalation timelines in some cases

🐛 We'll no longer suggest null/undefined JSON paths as part of alert attribute suggestions

🐛 We'll prompt you to upgrade users to an on-call seat when importing schedules from PagerDuty that are part of escalation paths