Healthchecks

To check the health status of your Metamapper instance, you can simply access the endpoint /health. It will return the following JSON object:

{
    "metastore": {
        "status": "healthly"
    },
    "scheduler": {
        "status": "healthy",
        "latest_scheduler_heartbeat": "2020-07-21 05:23:19+00:00"
    },
    "worker": {
        "status": "healthly"
    }
}

The status of each component can be healthy or unhealthy.

Metastore

The status of metastore depends on whether a valid connection could be made to the database.

Scheduler

The status of scheduler depends on when the latest scheduler heartbeat was received.

Every 30 seconds, the beat scheduler updates a database record with the current timestamp. If the last heatbeat was received more than 60 seconds ago, we consider the scheduler unhealthy.

Worker

The status of worker is determined by sending a small task through the default queue and confirming that it was processed via the result_backend.

If you have decided to not configure the result_backend or are using a broker that does not support the result_backend (such as Amazon SQS), the status will be set to not_configured.

Webserver

If the healthcheck endpoint does not return a 200 - OK status code, you can assume that the webserver and GraphQL endpoints are unhealthy.