Coming soon
Let an alert start the investigation
An issue or metric alert spawns an agent that pulls the event context, works the failure in a sandbox, and comments its diagnosis back on the Sentry issue.
Coming soon
An issue or metric alert spawns an agent that pulls the event context, works the failure in a sandbox, and comments its diagnosis back on the Sentry issue.
The Sentry integration is not released yet. Here is the shape it takes when it lands; email us if you want it sooner.
An issue alert or a metric alert starts an agent, the same way a pull request or a Linear issue does today.
The trigger names the Sentry projects it watches, so a noisy project never wakes an agent that does not cover it.
The agent starts with the event: the stack trace, the breadcrumbs, and the release it first appeared in.
The investigation happens in a sandbox with your repositories cloned and your toolchain installed, so the agent can reproduce rather than speculate.
The diagnosis lands as a comment on the Sentry issue, where whoever is on call is already looking.
Scoped credentials, hard budget caps, and a full transcript, exactly like every other session.