Errors

Hud sees every call to every function, so it catches errors as they happen, including exceptions your code catches. Each issue names the function where the failure started, not just the endpoint that returned the error.

The Errors rule

Flags failures and names the function that caused them:

• Endpoints: 5xx responses, and flows marked as failed.
• Queues: failures processing a message.
• Custom flows: failures in the flow.

Hud looks at 10-minute windows. By default it opens an issue as soon as there's an error, and alerts you when:

New Exception Type

Flags a function the first time it throws an exception type it has never thrown before, even if your code catches it. You hear about a new failure mode on its first occurrence, even if nothing has returned an error yet.

It only applies to services and functions Hud has tracked for at least 30 days.

Bad Request Errors

Flags 4xx responses on endpoints, with the function that caused them. Many 4xx responses are expected, like a request with a missing field, so by default Hud only alerts when at least 50% of an endpoint's requests in a 10-minute window return 4xx, with at least 50 of them. It starts turned off. Turn this rule on or off on the Rules page.

Errors after a rollout

Post-Deploy Error flags new errors, or a sharp rise in existing ones, on the endpoints, queues and custom flows a new version changed. See Post-Deploy Issues.

Failures that don't throw

Some failures don't throw or return an error status, like an endpoint that returns 200 OK with an error in the body. Mark these flows as failed in your code so Hud counts them. A 4xx response you mark as failed counts as an error even with Bad Request Errors off. See how in Node.js, Python and Java.

Change the defaults

To change these defaults for a service or an endpoint, open the rule on the Rules page. Detection conditions decide when Hud opens an issue, and alert conditions decide when it alerts you. While an issue keeps recurring, Hud alerts on it again every 3 days by default.