v139 · miscellaneous

Crash Reporting API: Specify crash-reporting to receive only crash reports

This feature ensures developers receive only crash reports by specifying the endpoint named `crash-reporting `. By default, crash reports are delivered to the `default` endpoint which receives many other kinds of reports besides crash reports. Developers can supply a separate URL to the well-known endpoint named `crash-reporting`, to dire

concepts

  1. Report-To Config

    A new Reporting endpoint group dedicated to crash reports. Sites can route crash dumps without also receiving deprecation, intervention, or CSP reports — easier to triage, smaller blast radius for an endpoint outage.

  2. Triage Dashboard

    Toggle between the old single-endpoint setup and the new dual-endpoint v139 setup. Emit a batch of mixed reports and watch how the noise ratio at your on-call pager drops.

  3. Reporting-Endpoints Header Builder

    Dial in endpoint URLs, toggle whether CSP uses the named report-to endpoint, decide whether to mirror crashes to default. The serialized header updates live and a simulated report mix routes into three inboxes you can watch fill up.

  4. Crash Report Inspector

    A simulated report stream of 6 types — crash, CSP violation, deprecation, intervention, permissions-policy. Toggle "Crash only" mode to see non-crash cards dimmed with a "filtered" badge. The live Report-To JSON header updates per filter mode with a copy button.

why it shipped

Developers report needing to receive crash reports exclusively. Currently, all crash reports are delivered to the endpoint named default with no override option. Developers are forced to receive not only crash reports but also unrelated reports like `Deprecated feature usage`.

references