# Check service status (/en/realtime-media/transcoding/build/integration)

> For AI agents: see the complete documentation index at [llms.txt](/llms.txt).

This page explains how to verify that your cloud transcoding service starts successfully and continues running properly.

## Check startup status [#check-startup-status]

Follow these steps to verify that the transcoding service has started successfully:

### Call `Create` immediately after you `Acquire` a `builderToken` [#call-create-immediately-after-you-acquire-a-buildertoken]

Ensure that you send a [`Create`](/en/api-reference/cloud-transcoding/restful) request within 2 seconds of each [`Acquire`](/en/api-reference/cloud-transcoding/restful) request. This ensures validity of the `builderToken`.

<CalloutContainer type="info">
  <CalloutDescription>
    If you acquire `builderToken`s in bulk and send `Create` requests in batches, the requests may fail.
  </CalloutDescription>
</CalloutContainer>

### Check the `Create` request response [#check-the-create-request-response]

A successful `Create` request returns an HTTP status code of `200`. If the status code is not `200`, handle it based on the returned code:

* **`206` (Partial Success):** Request timed out. Use a backoff strategy before retrying:

  * Wait 3 seconds, then retry
  * Wait 6 seconds, then retry again
  * Wait 9 seconds, then retry a third time
    If all retries fail, generate a new UID, call `Acquire` again to get a new `builderToken`, and retry `Create`.

* **`409` (Conflict):** The transcoding task has already started. No retry is needed.

* **`40x` (except `409`):** Parameter error. Check and correct the request parameters.

* **`50x` (Server Error):** Temporary server issue. Retry using the backoff strategy above. If all retries fail, change the UID and re-initiate the process.

* **Error code `65`:** Retry `Create` with the same parameters. Use a backoff strategy:

  * Retry after 3 seconds
  * Retry again after 6 seconds

### Set the appropriate `idleTimeout` [#set-the-appropriate-idletimeout]

The `idleTimeout` parameter in the `Create` request defaults to 300 seconds. This defines how long the transcoder stays active after all hosts leave the channel.

If hosts may temporarily leave and rejoin, set a larger `idleTimeout` to prevent the transcoder from shutting down too early. This ensures stable operation and avoids frequent restarts.

### Query the startup status [#query-the-startup-status]

After receiving a `taskId` from a successful `Create` request, wait 5 seconds and then call the [`Query`](/en/api-reference/cloud-transcoding/restful) method using a backoff strategy. Make sure the backoff interval is shorter than the `idleTimeout`.

* If the Query response shows status as `"STARTED"` or `"IN_PROGRESS"`, the transcoder has successfully started.
* If the `status` is still not `"STARTED"` or `"IN_PROGRESS"` 90 seconds after receiving the `taskId`, consider the startup failed due to timeout.

<CalloutContainer type="warning">
  <CalloutDescription>
    When retrying:

    * UIDs in the same channel must be unique. Prepare a backup UID for the transcoder in case you need to restart the task.
    * Alternate between the primary and backup UIDs as needed.
  </CalloutDescription>
</CalloutContainer>

## Check service running status [#check-service-running-status]

To confirm that the transcoder is still running, periodically call the `Query` method. If you need high reliability, Agora recommends calling `Query` every 2 minutes and taking appropriate action based on the HTTP response:

* **`40x` (Parameter Error):** Check and fix the request parameters.

* **`404` (Not Found):** If parameters are correct, this means the transcoder never started or exited prematurely. Retry `Query` using a backoff strategy (For example, 5s, 10s, 15s intervals) for confirmation.

* **`50x` (Server Error):** Indicates a temporary error, not necessarily that the transcoder has exited. Retry using increasing intervals (For example, 5s, 10s, 15s, 30s).

For further information, see [Status values](#status-values).

## Reference [#reference]

This section contains content that completes the information on this page, or points you to documentation that explains other aspects to this product.

### Status values [#status-values]

When querying the service status through the Go `Query` method, refer to the `services.cloudTranscoder.status` field in the response body.

| Service Status              | Description                             |
| --------------------------- | --------------------------------------- |
| `"serviceIdle"`             | Service has not started                 |
| `"serviceReady"`            | Service is ready to start               |
| `"serviceStarted"`          | Service has started                     |
| `"serviceInProgress"`       | Service is currently running            |
| `"serviceCompleted"`        | Service stopped, all tasks completed    |
| `"servicePartialCompleted"` | Service stopped with partial completion |
| `"serviceValidationFailed"` | Parameter validation failed             |
| `"serviceAbnormal"`         | Service exited unexpectedly             |
| `"serviceUnknown"`          | Service status is unknown               |