health-check.md

Health Checks for Upstream

Health Check of APISIX is based on lua-resty-healthcheck, you can use it for upstream.

Note that we only start the health check when the upstream is hit by a request. There won't be any health check if an upstream is configured but isn't in used.

The following is an example of health check:

curl http://127.0.0.1:9080/apisix/admin/routes/1 -H 'X-API-KEY: edd1c9f034335f136f87ad84b625c8f1' -X PUT -d '
{
    "uri": "/index.html",
    "plugins": {
        "limit-count": {
            "count": 2,
            "time_window": 60,
            "rejected_code": 503,
            "key": "remote_addr"
        }
    },
    "upstream": {
         "nodes": {
            "127.0.0.1:1980": 1,
            "127.0.0.1:1970": 1
        },
        "type": "roundrobin",
        "retries": 2,
        "checks": {
            "active": {
                "timeout": 5,
                "http_path": "/status",
                "host": "foo.com",
                "healthy": {
                    "interval": 2,
                    "successes": 1
                },
                "unhealthy": {
                    "interval": 1,
                    "http_failures": 2
                },
                "req_headers": ["User-Agent: curl/7.29.0"]
            },
            "passive": {
                "healthy": {
                    "http_statuses": [200, 201],
                    "successes": 3
                },
                "unhealthy": {
                    "http_statuses": [500],
                    "http_failures": 3,
                    "tcp_failures": 3
                }
            }
        }
    }
}'

The configures in checks are belong to health check, the type of checks contains: active or passive.

• active: To enable active health checks, you need to specify the configuration items under checks.active in the Upstream object configuration.

  • active.timeout: Socket timeout for active checks (in seconds), support decimals. For example 1.01 means 1010 milliseconds, 2 means 2000 milliseconds.

  • active.http_path: The HTTP GET request path used to detect if the upstream is healthy.

  • active.host: The HTTP request host used to detect if the upstream is healthy.

  • active.port: The customize health check host port (optional), this will override the port in the upstream node.

  The threshold fields of healthy are:

  • active.healthy.interval: Interval between health checks for healthy targets (in seconds), the minimum is 1.
  • active.healthy.successes: The number of success times to determine the target is healthy, the minimum is 1.

  The threshold fields of unhealthy are:

  • active.unhealthy.interval: Interval between health checks for unhealthy targets (in seconds), the minimum is 1.
  • active.unhealthy.http_failures: The number of http failures times to determine the target is unhealthy, the minimum is 1.
  • active.req_headers: Additional request headers. Array format, so you can fill in multiple headers.

• passive: To enable passive health checks, you need to specify the configuration items under checks.passive in the Upstream object configuration.

  The threshold fields of healthy are:

  • passive.healthy.http_statuses: If the current response code is equal to any of these, set the upstream node to the healthy state. Otherwise ignore this request.
  • passive.healthy.successes: Number of successes in proxied traffic (as defined by passive.healthy.http_statuses) to consider a target healthy, as observed by passive health checks.

  The threshold fields of unhealthy are:

  • passive.unhealthy.http_statuses: If the current response code is equal to any of these, set the upstream node to the unhealthy state. Otherwise ignore this request.
  • passive.unhealthy.tcp_failures: Number of TCP failures in proxied traffic to consider a target unhealthy, as observed by passive health checks.
  • passive.unhealthy.timeouts: Number of timeouts in proxied traffic to consider a target unhealthy, as observed by passive health checks.
  • passive.unhealthy.http_failures: Number of HTTP failures in proxied traffic (as defined by passive.unhealthy.http_statuses) to consider a target unhealthy, as observed by passive health checks.

The health check status can be fetched via GET /v1/healthcheck in control API.