> ## Documentation Index
> Fetch the complete documentation index at: https://docs.paradime.io/llms.txt
> Use this file to discover all available pages before exploring further.

# GCP Dataflow CLI

> Launch Google Cloud Dataflow jobs from classic or Flex templates via Paradime Bolt using the paradime run gcp-dataflow CLI commands with run monitoring.

The Paradime SDK provides a CLI command to interact with Google Cloud Dataflow, allowing you to launch a Dataflow job from a classic or flex template.

<Warning>
  🔑 **Authentication Required**

  You will need a GCP service account JSON key file with permissions to launch Dataflow jobs in your project. Provide it via the `--service-account-key-file` flag or the `GCP_SERVICE_ACCOUNT_KEY_FILE` environment variable.
</Warning>

## Launch a Dataflow Job

Launch a Dataflow job from a template (classic or flex), optionally passing template parameters and waiting for completion.

### CLI Command

```bash theme={"system"}
paradime run gcp-dataflow-trigger
```

#### Options

| Flag                         | Type              | Description                                                                                                                                                   |
| ---------------------------- | ----------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `--service-account-key-file` | Required, TEXT    | Path to your GCP service account JSON key file. Can be set via `GCP_SERVICE_ACCOUNT_KEY_FILE` environment variable.                                           |
| `--project`                  | Required, TEXT    | Your GCP project ID. Can be set via `GCP_PROJECT_ID` environment variable.                                                                                    |
| `--location`                 | Required, TEXT    | GCP region (e.g. 'us-central1'). Can be set via `GCP_LOCATION` environment variable.                                                                          |
| `--template-path`            | Required, TEXT    | GCS path to the Dataflow template (e.g. gs\://bucket/templates/my-template).                                                                                  |
| `--job-name`                 | Required, TEXT    | Name for the launched Dataflow job.                                                                                                                           |
| `--template-type`            | Optional, CHOICE  | Type of Dataflow template (classic or flex). Choices: `classic`, `flex`. Default: `classic`.                                                                  |
| `--parameters`               | Optional, TEXT    | JSON string of template parameters (e.g. `{"inputFile": "gs://..."}`).                                                                                        |
| `--wait` / `--no-wait`       | Optional, Flag    | Wait for the Dataflow job to complete before returning. Default: `--wait`. Can be set via `GCP_DATAFLOW_TRIGGER_WAIT` environment variable.                   |
| `--timeout`                  | Optional, INTEGER | Maximum time to wait for completion (in minutes). Only used with `--wait`. Default: 1440. Can be set via `GCP_DATAFLOW_TRIGGER_TIMEOUT` environment variable. |
| `--help`                     | Optional, Flag    | Show the help message and exit.                                                                                                                               |

<Info>
  **Recommended Setup**

  For security and convenience, set your GCP credentials as environment variables:

  ```bash theme={"system"}
  GCP_SERVICE_ACCOUNT_KEY_FILE="/path/to/service-account-key.json"
  GCP_PROJECT_ID="your_project_id"
  GCP_LOCATION="us-central1"
  ```
</Info>

### Usage Examples

```bash theme={"system"}
# Launch a classic template job (using environment variables)
paradime run gcp-dataflow-trigger \
  --template-path "gs://bucket/templates/my-template" \
  --job-name "my-dataflow-job"

# Launch a flex template job with parameters
paradime run gcp-dataflow-trigger \
  --template-path "gs://bucket/templates/my-flex-template" \
  --job-name "my-flex-job" \
  --template-type flex \
  --parameters '{"inputFile": "gs://bucket/input.csv"}'

# With explicit credentials and no wait
paradime run gcp-dataflow-trigger \
  --service-account-key-file "/path/to/key.json" \
  --project "your_project_id" \
  --location "us-central1" \
  --template-path "gs://bucket/templates/my-template" \
  --job-name "my-dataflow-job" \
  --no-wait
```

## Environment Variable Reference

| Environment Variable           | Description                                            |
| ------------------------------ | ------------------------------------------------------ |
| `GCP_SERVICE_ACCOUNT_KEY_FILE` | Path to your GCP service account JSON key file         |
| `GCP_PROJECT_ID`               | Your GCP project ID                                    |
| `GCP_LOCATION`                 | GCP region (e.g. 'us-central1')                        |
| `GCP_DATAFLOW_TRIGGER_WAIT`    | Wait for the Dataflow job to complete before returning |
| `GCP_DATAFLOW_TRIGGER_TIMEOUT` | Maximum time to wait for completion (in minutes)       |


## Related topics

- [GCP Dataproc CLI](/developers/paradime-cli/gcp-dataproc-cli.md)
- [GCP Datastream CLI](/developers/paradime-cli/gcp-datastream-cli.md)
- [GCP Cloud Function CLI](/developers/paradime-cli/gcp-cloud-function-cli.md)
- [GCP Cloud Run CLI](/developers/paradime-cli/gcp-cloud-run-cli.md)
- [GCP BigQuery Data Transfer CLI](/developers/paradime-cli/gcp-bigquery-transfer-cli.md)