*Issue:* During service disruptions, there was no visibility into the health of the database connection pools, making it difficult to distinguish between credential failures and connection exhaustion.
*Corrective Action:* Implement Prometheus metrics to track active versus idle database connections. A new alert should be configured to trigger when active connections exceed 90% of the available pool for a sustained period.
*Result:* This will provide engineers with immediate insight into database connection health, enabling faster diagnosis and proactive response before the service becomes unavailable.