Background
Operator triggers a Redis Pod restart for every user-defined Redis configuration change, regardless of the configuration item type. While this approach ensures configuration effectiveness, it introduces inefficiencies and potential availability concerns.
Problem Description
Unnecessary Service Interruptions: Many Redis configuration directives (e.g., maxmemory-policy, slowlog-log-slower-than) are designed to be dynamically modified at runtime using the CONFIG SET command without requiring a server restart. The current operator's approach of restarting Pods for all configuration changes leads to unnecessary service interruptions, reducing Redis cluster availability.
Frequent Pod restarts consume extra resources and time, decreasing the efficiency of configuration management, especially in scenarios with frequent configuration updates.