GET Healthcheck

Healthcheck return details on the health of Kafka and its topics.

Request

GET /_/backbeat/api/healthcheck

Responses

If the response is not an HTTP error, it is structured as an object with two main keys: topics and internalConnections.

The topics key returns details on the Kafka CRR topic only. The name field returns the Kafka topic name, and the partitions field returns details of each partition for the CRR topic. id carries the partition ID, leader contains the current node responsible for all reads and writes for the given partition, the replicas array is the list of nodes that replicate the log for the given partition, and the isrs array is the list of in-sync replicas.

topics: {
  <topicName>: {
    name: <string>,
    partitions: {
      id: <number>,
      leader: <number>,
      replicas: [<number>, ...],
      isrs: [<number>, ...],
    },
    ...
  },
}

The internalConnections key returns general details on the health of the system as a whole. It contains an object of three keys:

• isrHealth, which carries a value of either ok or error.
• zookeeper, which shows a status and status details (see https://github.com/alexguan/node-zookeeper-client#state for more detail).
• kafkaProducer, which carries a value of either ok or error, and reports the health of all producers for every topic.

isrHealth reports whether the minimum in-sync replicas count for every partition is met. The zookeeper field indicates whether ZooKeeper is running properly. The kafkaProducer field reports the health of all Kafka producers for every topic.

The zookeeper:details field provides a status name and status code provided directly from the node-zookeeper), library.

internalConnections: {
  isrHealth: <ok || error>,
  zookeeper: {
    status: <ok || error>,
    details: {
      name: <value>,
      code: <value>
    }
  },
  kafkaProducer: {
    status: <ok || error>
  }
}

Example

Note

The following example is redacted for brevity.

{
  "topics": {
    "backbeat-replication": {
      "name": "backbeat-replication",
      "partitions": [
        {
          "id": 2,
          "leader": 4,
          "replicas": [2,3,4],
          "isrs": [4,2,3]
        },
        ...
        {
          "id": 0,
          "leader": 2,
          "replicas": [0,1,2],
          "isrs": [2,0,1]
        }
      ]
    }
  },
  "internalConnections": {
    "isrHealth": "ok",
    "zookeeper": {
      "status": "ok",
      "details": {
        "name": "SYNC_CONNECTED",
        "code": 3
      }
    },
    "kafkaProducer": {
      "status": "ok"
    }
  }
}

Healthcheck Details

Backbeat healthchecks returns an object with two main keys: topics and internalConnections.

The topics key returns details on the Kafka CRR topic only. The name field returns the Kafka topic name, and the partitions field returns details of each partition for the CRR topic. The id is the partition id, the leader is the current node responsible for all reads and writes for the given partition, the replicas array is the list of nodes that replicate the log for the given partition, and the isrs array is the list of in-sync replicas.

topics: {
  <topicName>: {
    name: <value>,
      partitions: [
        {
          id: <value>,
          leader: <value>,
          replicas: [<value>, ...],
          isrs: [<value>, ...]
        },
      ...
    ]
  }
}

The internalConnections key returns general details on the health of the system as a whole. isrHealth checks if the minimum in-sync replicas for every partition is met, the zookeeper field checks if ZooKeeper is running properly, and the kafkaProducer field checks the health of all Kafka producers for every topic.

healthcheck/deep

Deep healthchecks return the health of every available partition for the replication topic.

Rather than getting an internal status variable or calling an internal status function to check the health of the replication topic, a deep healthcheck produces and consumes a message directly to the replication topic for every available partition.

If there is no HTTP response error, the response is a JSON-structured object of:

topicPartition: <ok || error>,
...
timeElapsed: <value>

Example Output

{
    "0":"ok",
    "1":"ok",
    "2":"error",
    "3":"ok",
    "4":"ok",
    "timeElapsed":560
}