"/health/service/X?passing" is missing services

[mmeinardi]

Sometimes requesting a list of healthy services (by using ?passing parameter) is missing the ones that pass all the health checks.

Example:

Query 1:
http://localhost:8500/v1/health/service/some-service

{
    "Checks": [
        {
            "CheckID": "service:92bcd373e541",
            "Name": "Service 'some-service' check",
            "Node": "consul-192.168.3.223",
            "Notes": "",
            "Output": "",
            "ServiceID": "92bcd373e541",
            "ServiceName": "some-service",
            "Status": "passing"
        },
        {
            "CheckID": "serfHealth",
            "Name": "Serf Health Status",
            "Node": "consul-192.168.3.223",
            "Notes": "",
            "Output": "Agent not live or unreachable",
            "ServiceID": "",
            "ServiceName": "",
            "Status": "critical"
        }
    ],
    "Node": {
        "Address": "192.168.3.223",
        "Node": "consul-192.168.3.223"
    },
    "Service": {
        "ID": "92bcd373e541",
        "Port": 4001,
        "Service": "some-service"
    }
},
{
    "Checks": [
        {
            "CheckID": "service:4a1c7d1b374f",
            "Name": "Service 'some-service' check",
            "Node": "consul-192.168.3.165",
            "Notes": "",
            "Output": "Exception: () \n",
            "ServiceID": "4a1c7d1b374f",
            "ServiceName": "some-service",
            "Status": "critical"
        },
        {
            "CheckID": "serfHealth",
            "Name": "Serf Health Status",
            "Node": "consul-192.168.3.165",
            "Notes": "",
            "Output": "Agent not live or unreachable",
            "ServiceID": "",
            "ServiceName": "",
            "Status": "critical"
        }
    ],
    "Node": {
        "Address": "192.168.3.165",
        "Node": "consul-192.168.3.165"
    },
    "Service": {
        "ID": "4a1c7d1b374f",
        "Port": 49154,
        "Service": "some-service"
    }
},
{
    "Checks": [
        {
            "CheckID": "service:e64e53ec404f",
            "Name": "Service 'some-service' check",
            "Node": "consul-192.168.3.129",
            "Notes": "",
            "Output": "",
            "ServiceID": "e64e53ec404f",
            "ServiceName": "some-service",
            "Status": "passing"
        },
        {
            "CheckID": "serfHealth",
            "Name": "Serf Health Status",
            "Node": "consul-192.168.3.129",
            "Notes": "",
            "Output": "Agent alive and reachable",
            "ServiceID": "",
            "ServiceName": "",
            "Status": "passing"
        }
    ],
    "Node": {
        "Address": "192.168.3.129",
        "Node": "consul-192.168.3.129"
    },
    "Service": {
        "ID": "e64e53ec404f",
        "Port": 4001,
        "Service": "some-service"
    }
}

Query 2:
http://localhost:8500/v1/health/service/some-service?passing

[]

But the 3rd service is clearly ok.

The bug seems to exist in this function:
https://github.com/hashicorp/consul/blob/master/command/agent/health_endpoint.go#L118
After finding out that one of the health check is not passing we should remove the service from the output and break - moving on to next services.

[armon]

Good catch! Thanks!