lb/Latency

This runbook diagnoses and troubleshoots latency issues with Application Load Balancers.

Product: Load balancing Kind: Debugging Tree

Description

It analyzes key metrics to identify potential bottlenecks and performance problems.

Key Investigation Areas:

  • Backend Latency:
    • Measures the time taken for backends to respond to requests, checking if it exceeds a configurable threshold.
  • Request Count Per Second (QPS):
    • Monitors the rate of incoming requests to the load balancer, checking if it exceeds a configurable threshold. A high request count coupled with high latency might suggest overload.
  • 5xx Error Rate:
    • Calculates the percentage of 5xx server errors, indicating problems on the backend servers. This check uses a configurable threshold and considers the request count to provide a meaningful error rate.

Executing this runbook

gcpdiag runbook lb/latency \
  -p project_id=value \
  -p forwarding_rule_name=value \
  -p region=value \
  -p backend_latency_threshold=value \
  -p request_count_threshold=value \
  -p error_rate_threshold=value

Parameters

Name Required Default Type Help
project_id True None str The Project ID where the load balancer is located
forwarding_rule_name True None str The name of the forwarding rule associated with the Load Balancer to check
region False None str The region where the forwarding rule is located
backend_latency_threshold False None float Threshold for backend latency in milliseconds.
request_count_threshold False None float Threshold for average request count per second.
error_rate_threshold False None float Threshold for error rate (percentage of 5xx errors).

Get help on available commands

gcpdiag runbook --help

Potential Steps