google_compute_health_check cost estimation

A google_compute_health_check defines how a load balancer or managed instance group probes its backends: the protocol (HTTP, HTTPS, TCP, gRPC), port, interval, timeout, and healthy/unhealthy thresholds. Health checks are free. There is no charge for the check, for the frequency of probes, or for the number of backends probed.

Because the check is free, the cost you should think about is what it is attached to. A health check feeds a backend service, and the Cloud Load Balancing charge lands on the forwarding rule (hourly plus data processed), not the check. A health check attached to a managed instance group drives autohealing, recreating unhealthy VMs, and those VMs carry the compute cost. The check is the probe, not the meter.

One practical note: probe traffic itself is not billed, and Google's health-check prober IP ranges are documented so you can allow them in firewall rules for free. c3x prices the forwarding rule and the backends a health check is associated with and treats the health check as the free liveness config it is.

Terraform example

A minimal but realistic configuration that C3X can estimate.

resource "google_compute_health_check" "app" {
  name                = "app-hc"
  check_interval_sec  = 5
  timeout_sec         = 5
  healthy_threshold   = 2
  unhealthy_threshold = 3

  http_health_check {
    port         = 8080
    request_path = "/healthz"
  }
}

Pricing dimensions

What you actually pay for when you provision google_compute_health_check.

Dimension	Unit	What's being charged
Health check	free	The health check, its probes, and the number of backends probed have no charge. $0
Forwarding rule it feeds	per hour + per GB	The load balancer the health check informs bills on the forwarding rule plus data processed. ~$0.025/hour per rule + data
Backends it protects	per hour	The VMs behind the check, including replacements autohealing creates, carry their own cost.

Optimization tips

Common ways to reduce google_compute_health_check cost without changing the workload.

Allow prober ranges without extra cost

Google's health-check prober IP ranges are fixed and documented. Allowing them in a firewall rule is free and avoids false-unhealthy backends that would trigger costly recreations.

Tune thresholds to avoid churn

Aggressive unhealthy thresholds can flap backends in and out, and with autohealing that recreates VMs unnecessarily. Sensible thresholds keep the paid fleet stable.

Share one health check across backends

A single health check definition can serve multiple backend services. There is no cost saving on the free check, but it keeps config simple and consistent.

FAQ

Does a health check cost money?

No. Health checks are free, including the probe traffic and the number of backends probed. Cost sits on the load balancer's forwarding rule and on the backends the check protects.

Is health-check probe traffic billed?

No. Probe traffic from Google's health-check prober ranges is not billed, and those ranges are documented so you can allow them in firewall rules for free.

How does a health check affect my bill?

Indirectly. It decides which backends receive traffic and, when attached to a managed instance group, triggers autohealing that recreates VMs. Those VMs bill; the check itself is free.