Database Connection Pool Health Probe with Adaptive Max-Connection Sizing

[JamesEjembi]

Problem Statement

Database connection pools are configured with static max-connection limits. Under varying load, pools either exhaust connections during spikes or waste memory during lulls. Implement an adaptive connection pool health probe that monitors query latency distribution, connection wait times, and pool utilization, then dynamically adjusts max-connections.

Technical Bounds

• Probe interval: 10s
• Min connections: 5, Max: 200
• Adjustment step: +/- 5 per interval
• Latency threshold: p95 < 100ms before scaling down
• Wait threshold: p90 < 50ms before scaling up
• Cooldown: 60s between consecutive same-direction adjustments

Steps

1. Create health probe collecting pool metrics
2. Implement PID-controller adjustment logic
3. Add Prometheus gauges for current/target pool sizes
4. Integration test verifying auto-scaling under load spike

[ekwe7]

I would love to work on this issue. I have experience building production ready applications and collaborating through GitHub workflows. I'll carefully analyze the current implementation, propose a reliable solution, and submit a well-tested PR. My focus is on clean code, maintainability, and timely delivery