Airflow

A small PHP client for the Apache Airflow HTTP REST API.

Airflow versions

The client defaults to the Airflow 2.x REST API base path /api/v1. Airflow 3.x serves the stable API under /api/v2; set api_path accordingly:

'api_path' => '/api/v2',

Quick start

use Erikwang2013\BigData\BigData;

$airflow = BigData::driver('airflow', [
    'baseUrl' => 'http://127.0.0.1:8080',
    'username' => 'airflow',
    'password' => 'airflow',
]);

foreach ($airflow->dags()['dags'] as $dag) {
    echo $dag['dag_id'], PHP_EOL;
}

Direct instantiation:

use Erikwang2013\BigData\Driver\Airflow\AirflowClient;
use Erikwang2013\BigData\Driver\Airflow\AirflowConfig;

$airflow = new AirflowClient(AirflowConfig::fromArray([
    'baseUrl' => 'http://127.0.0.1:8080',
    'username' => 'airflow',
    'password' => 'airflow',
]));
// Optional HTTP client and PSR-3 logger: new AirflowClient($config, $http, $logger)

A bearer token (e.g. from a JWT-auth setup) is used instead of basic auth:

'baseUrl' => 'https://airflow.example.com',
'token' => 'eyJhbGciOi...',

Configuration

AirflowConfig::fromArray() accepts:

API

$airflow->dags(int $limit = 100, int $offset = 0);                // GET /dags
$airflow->dag(string $dagId);                                     // GET /dags/{dagId}
$airflow->triggerDag(string $dagId, array $conf = [], ?string $dagRunId = null); // POST /dags/{dagId}/dagRuns
$airflow->dagRuns(string $dagId, int $limit = 100, int $offset = 0);             // GET /dags/{dagId}/dagRuns
$airflow->dagRun(string $dagId, string $dagRunId);                // GET /dags/{dagId}/dagRuns/{runId}
$airflow->taskInstances(string $dagId, string $dagRunId);         // GET /dags/{dagId}/dagRuns/{runId}/taskInstances
$airflow->taskLog(string $dagId, string $dagRunId, string $taskId, int $tryNumber = 1); // GET .../taskInstances/{taskId}/logs/{tryNumber}
$airflow->variables(int $limit = 100, int $offset = 0);           // GET /variables
$airflow->setVariable(string $key, string $value);                // POST /variables
$airflow->deleteVariable(string $key);                            // DELETE /variables/{key}

JSON endpoints return the decoded response body as an array. taskLog() returns the raw log text (Accept: text/plain).

Logging

AirflowClient accepts an optional PSR-3 logger as its third constructor argument and defaults to Psr\Log\NullLogger. Debug records contain the HTTP method and URI only — credentials are never logged.

Errors

Transport failures throw Erikwang2013\BigData\Exception\ConnectionException. HTTP status >= 400, invalid JSON, or an unreadable response raises Erikwang2013\BigData\Exception\RequestException, which exposes public $statusCode and $body when available; a 404 is reported the same way. Both extend Erikwang2013\BigData\Exception\BigDataException.

Notes

• baseUrl must not include the API path; use the api_path option for that.
• Pagination is limited to the limit/offset arguments exposed by each method.
• No automatic retries are performed.