---
title: "Job Failed"
method: POST
path: "webhookEvents_jobFailed"
---

# Job Failed

`POST webhookEvents_jobFailed` (webhook)

Sent when an indexing job fails. The job's final status is `failed`.

`data.error_code` says why. `all_files_failed` means every file failed and nothing from the job was indexed. `embedding_stalled` means the job's video or audio files never finished embedding, so nothing was indexed. `execution_failed` means the job stopped with an error, and `data.files.indexed` counts any files indexed before it stopped. Fetch the job at `data.url` for the per-file errors, then retry.

Each indexing job sends exactly one event, when it reaches its final status. See [Webhooks](/guides/webhooks) to add an endpoint and verify signatures.

## Headers

- `webhook-id` string, required
- `webhook-timestamp` string, required
- `webhook-signature` string, required

## Payload

- JobWebhookEvent — The body of every job webhook event. It carries ids, counts and timestamps only, never file names, file paths, error messages or `custom_metadata`.
  - `id` string, required — `evt_` followed by the job id. The same for every delivery of this job's event.
  - `type` 'job.completed' | 'job.completed_with_errors' | 'job.failed' | 'job.timed_out' | 'job.cancelled', required — The event name.
  - `timestamp` string, date-time, required — When the job finished, in ISO 8601 UTC.
  - `data` JobWebhookEventData, required — The indexing job the event reports on.
    - `job_id` string, required — The indexing job's id, as returned when the job was started.
    - `status` 'completed' | 'completed_with_errors' | 'failed' | 'timed_out' | 'cancelled', required — The job's final status. It matches the event name: `job.completed` carries `completed`, and so on.
    - `source` 'api' | 'sync', required — `api` when the job was started by a call to an index endpoint, `sync` when a storage sync started it.
    - `job_type` string, nullable — The kind of indexing job, for example `index_s3_directory`.
    - `collection_id` string, nullable — The id of the collection the job indexed into.
    - `collection_name` string, nullable — The collection's name. `null` when an endpoint receiving the event has `include_collection_name` set to `false`.
    - `environment` string, required — The environment the job ran in: `development`, `staging` or `production`. Endpoints belong to the organization and receive jobs from every environment.
    - `sync_id` string, nullable — The sync that started the job, or `null` for jobs started through the API.
    - `created_at` string, date-time, nullable — When the job was created, in ISO 8601 UTC.
    - `completed_at` string, date-time, nullable — When the job finished, in ISO 8601 UTC.
    - `files` JobWebhookFiles, required — File counts for the job when it finished.
      - `total` integer, required — Files in the job.
      - `indexed` integer, required — Files indexed.
      - `failed` integer, required — Files that failed.
      - `skipped` integer, required — Files skipped, for example because `skip_existing` matched an existing document.
    - `error_code` string, nullable — Why the job did not fully succeed. Known values are `all_files_failed`, `execution_failed`, `timed_out` and `embedding_stalled`. `null` when the job completed. New values may be added, so treat an unknown value as a failure rather than rejecting the event.
    - `test` boolean, required — `true` for a test event, `false` for a real job.
    - `url` string, uri, required — The job's status URL on the Captain API. Fetch it for per-file detail.

## Acknowledgement `200`

Webhook received successfully

---

[API](https://skmtc.dev/runcaptain/apis/api-reference.md) · [All operations](https://skmtc.dev/runcaptain/apis/api-reference/llms.txt) · [OpenAPI document](https://skmtc.dev/runcaptain/apis/api-reference/revisions/ac61e472bb7d?raw)