## Healthcheck Probe

MinIO server has two healthcheck related un-authenticated endpoints, a liveness probe to indicate if server is responding, cluster probe to check if server can be taken down for maintenance.

- Liveness probe available at `/minio/health/live`
- Cluster probe available at `/minio/health/cluster`

# MinIO Healthcheck

MinIO server exposes three un-authenticated, healthcheck endpoints liveness probe and a cluster probe at `/minio/health/live` and `/minio/health/cluster` respectively.

## Liveness probe

This probe always responds with '200 OK'. Only fails if 'etcd' is configured and unreachable. When liveness probe fails, Kubernetes like platforms restart the container.

```
livenessProbe:
  httpGet:
    path: /minio/health/live
    port: 9000

      #
      # If the probe is enabled, it is recommended to have delay=0s,period=15s,failureThreshold=4.
      # This ensures the pod is marked ready immediately after the startup probe passes (which has a 1s poll interval),
      # and doesn't spam the readiness endpoint too much
      #
      # If the probe is disabled, it is recommended to have delay=1s,period=2s,failureThreshold=30.

	HostProbeSNATIP = netip.MustParseAddr(env.RegisterStringVar("HOST_PROBE_SNAT_IP", DefaultHostProbeSNATIP, "").Get())
)

const (
	// to reliably identify kubelet healthprobes from inside the pod (versus standard kube-proxy traffic,
	// since the IP is normally the same), we SNAT identified host probes in the host netns to a fixed APIPA IP.
	//

        proxyMetadata:
          # this should override the value from the global meshconfig
          PROXY_CONFIG_ANNOT_VALUE: bar
    labels: {}
  template:
    ports: {}
    serviceAccount: vm-serviceaccount
  probe:
    httpGet:

      #
      # If the probe is enabled, it is recommended to have delay=0s,period=15s,failureThreshold=4.
      # This ensures the pod is marked ready immediately after the startup probe passes (which has a 1s poll interval),
      # and doesn't spam the readiness endpoint too much
      #
      # If the probe is disabled, it is recommended to have delay=1s,period=2s,failureThreshold=30.

type remoteVersionID string

// newWarmBackend instantiates the tier type specific WarmBackend, runs
// checkWarmBackend on it.
func newWarmBackend(ctx context.Context, tier madmin.TierConfig, probe bool) (d WarmBackend, err error) {
	switch tier.Type {
	case madmin.S3:
		d, err = newWarmBackendS3(*tier.S3, tier.Name)
	case madmin.Azure:
		d, err = newWarmBackendAzure(*tier.Azure, tier.Name)
	case madmin.GCS:

			// TCP starts sending keepalive probes
			_ = syscall.SetsockoptInt(fd, syscall.IPPROTO_TCP, syscall.TCP_KEEPIDLE, 15)

			// Number of probes.
			// ~ cat /proc/sys/net/ipv4/tcp_keepalive_probes (defaults to 9, we reduce it to 5)
			_ = syscall.SetsockoptInt(fd, syscall.IPPROTO_TCP, syscall.TCP_KEEPCNT, 5)

			// Wait time after successful probe in seconds.

	}

	return openNetns, nil
}

// AddPodToMesh adds a pod to mesh by
// 1. Getting the netns
// 2. Adding the pod's IPs to the hostnetns ipsets for node probe checks
// 3. Creating iptables rules inside the pod's netns
// 4. Notifying ztunnel via GRPC to create a proxy for the pod
//

kind: WorkloadGroup
metadata:
  name: foo
  namespace: bar
spec:
  metadata:
    annotations: {}
    labels: {}
  template:
    ports: {}
    serviceAccount: vm-serviceaccount
  probe:
    httpGet: