lb/Unhealthy Backends

Load Balancer Unhealthy Backends Analyzer.

Product: Load balancing Kind: Debugging Tree

Description

This runbook helps investigate why backends in a load balancer are unhealthy. It confirms and summarizes the current health status of the backends, aiding in identifying any unhealthy instances.

Key Investigation Areas:

  • Firewalls:
    • Verifies if firewall rules are properly configured to allow health check traffic.
  • Port Configuration:
    • Validates if the health check and serving ports are configured correctly, ensuring they are not mismatched.
  • Logging:
    • Checks if health check logging is enabled to aid in troubleshooting.
  • Health Check Logs (if enabled):
    • Analyzes the latest health check logs to identify the specific reasons for backend unhealthiness:
      • Timeouts: Identifies if the backend is timing out and provides potential causes and remediation steps.
      • Unhealthy: Indicates that the backend is reachable but doesn’t meet the health check’s criteria. It provides guidance on the expected health check behavior and suggests configuration checks.
      • Unknown: Explains the potential reasons for the “UNKNOWN” health state and suggests actions like adjusting timeouts or checking for Google Cloud outages.
  • Past Health Check Success:
    • Checks if the health check has worked successfully in the past to determine if the issue is recent or ongoing.

Executing this runbook

gcpdiag runbook lb/unhealthy-backends \
  -p project_id=value \
  -p backend_service_name=value \
  -p region=value

Parameters

Name Required Default Type Help
project_id True None str The Project ID of the resource under investigation
backend_service_name True None str The name of the backend service that you want to investigate
region False None str The region configured for the load balancer (backend service). If not provided, the backend service is assumed to be global.

Get help on available commands

gcpdiag runbook --help

Potential Steps