How-to

If you want to use the healthcheck feature for a container. You can inspect a container with docker inspect $CONTAINER_NAME.

The response is an array with one JSON object:

[
    {
        "Id": "CONTAINER_ID",
        "Created": "2021-11-09T21:18:05.552662745Z",
        "Path": "/start.py",
        "Args": [],
        "State": {
            "Status": "running",
            "Running": true,
            "Paused": false,
            "Restarting": false,
            "OOMKilled": false,
            "Dead": false,
            "Pid": 12872,
            "ExitCode": 0,
            "Error": "",
            "StartedAt": "2021-11-09T21:18:08.697693764Z",
            "FinishedAt": "0001-01-01T00:00:00Z",
            "Health": {
                "Status": "healthy",
                "FailingStreak": 0,
                "Log": [
                    {
                        "Start": "2021-11-10T18:51:32.122181664Z",
                        "End": "2021-11-10T18:51:32.304082149Z",
                        "ExitCode": 0,
                        "Output": "OK"
                    }
                ]
            }
        },
        "Image": "IMAGE_SH"
        "HostConfig": {
          ...
        },
        "GraphDriver": {
          ...
        },
        "Mounts": [
          ...
        ],
        "Config": {
            "Hostname": "hostname",
            "Domainname": "",
            "User": "",
            "AttachStdin": false,
            "AttachStdout": false,
            "AttachStderr": false,
            "ExposedPorts": {
                "8008/tcp": {},
                "8009/tcp": {},
                "8448/tcp": {}
            },
            "Tty": false,
            "OpenStdin": false,
            "StdinOnce": false,
            "Env": [
              ...
            ],
            "Cmd": null,
            "Healthcheck": {
                "Test": [
                    "CMD-SHELL",
                    "curl -fSs http://localhost:8008/health || exit 1"
                ],
                "Interval": 60000000000,
                "Timeout": 5000000000
            },
            "Image": "IMAGE_NAME",
            "Volumes": {
                "/data": {}
            },
            "WorkingDir": "",
            "Entrypoint": [
                "/start.py"
            ],
            "OnBuild": null,
            "Labels": {
              ...
            }
        },
        "NetworkSettings": {
          ...
        }
    }
]

If you take a look at the object, you will find the healthcheck cmd.

curl -fSs http://localhost:8008/health || exit 1

Check the curl docs, how the options return a nice result. Now what you can do, you can search for it using grep:

docker inspect $CONTAINER_NAME | grep "healthy"
                "Status": "healthy",

The problem in this case is, you want a boolean result. So you have different option, you can handle this case with option -o of the grep search:

docker inspect docker_synapse_1 | grep -o "healthy"
"healthy"

Okay cool, let’s check the exit code. echo $? will return 0, so know you can use this pipe in a shell script and handle if conditions. As you may know, we are not the only one with this problem. So Docker provides a convenient solution.

Docker inspect provides the option —format, which allows us to ask for a specified JSON structure. It can do more, but for our case it is enough.

docker inspect --format {% raw %}'{{.State.Health.Status }}{% endraw %}' $CONTAINER_NAME

The command above will return healthy. if you want the double quotes, you can add json after the curly brackets, like {% raw %}{{json .State...}}{% endraw %}. So there you have it, you can now check against this string using grep or just [[ $? != "healthy"]].

Advantages:

• No pipe required.
• Returning the exact value of the object key.
• Just Docker, no strings attached.