Capture & report errors

Use the Node or Bun client to report handled errors and finish sending events before short-lived work exits.

Capture exceptions

Rasputin captures uncaught errors automatically. Use captureException when you catch an error yourself but still want Rasputin to report it. It accepts any thrown value and never throws. Bare captureException(error) reports the exception only. To attach runtime recorder evidence, pass observation from scope.observeError(error).

try {
  await chargeCustomer();
} catch (error) {
  rasputin.captureException(error);
}

Send pending events before exit

Rasputin sends events in the background. A script or serverless function can finish before that happens, so call flush before it exits. Long-running servers flush automatically.

await rasputin.flush();

Client options

These are the fields you can pass to the RasputinInit client. Rasputin discovers source from package boundaries; root settings are optional.

Field	Required	What it does
`dsn`	Required	Public project DSN from project settings. Used only to send events.
`projectApiKey`	Required	Secret project API key for deploy announcement and instrumentation manifest upload.
`environment`	Required	Where the error happened, such as production or staging.
`release`	Required	The Git commit SHA for this deployment.
`moduleUrl`	Optional	Helps normalize local stack paths. Use import.meta.url.
`repoRoot`	Optional	A local filesystem display root. It does not control source identity.
`sourceRoot`	Optional	The Git folder for a single-package service when automatic matching is ambiguous.
`sourceRoots`	Optional	Package-name to Git-folder mappings for an ambiguous monorepo.
`enabled`	Optional	Turns reporting on or off. Defaults to true.
`logSuccess`	Optional	Prints a configuration summary when setup succeeds. Defaults to true.
`executionRecorder`	Optional	Optional. Does not change basic error reporting. Enables the runtime recorder so errors can include arguments, return values, and call history from the failing work.