Introduction

Monitors Redis response latency by sending PING commands and fails the experiment if the response time exceeds the configured threshold.

This check is designed to be combined with attacks that affect Redis responsiveness, such as Pause Clients or network-level fault injections. For monitoring connection counts, use Connection Count Check instead.

Use Cases

• Verify that Redis response latency stays within SLO bounds during a Pause Clients attack with WRITE mode (reads should remain fast).
• Detect latency spikes during chaos experiments that inject network latency or CPU stress on the Redis host.
• Set a latency guardrail to fail experiments if Redis becomes too slow, preventing cascading failures during testing.

Parameters

Parameter	Description	Default
Duration	How long the check runs. Latency is measured continuously by sending PING commands for this duration.	60s
Max Latency	Maximum allowed response latency in milliseconds. If any PING response exceeds this threshold, the experiment fails.	100