Server Health Check Flapping Indicates Network Instability
warningreliabilityUpdated Feb 23, 2026
Frequent status changes between UP and DOWN for backend servers, visible in haproxy_server_status transitions, indicates network instability or overly aggressive health check timeouts.
How to detect:
Track the rate of change in haproxy_server_status over time windows. Monitor Downtime Check counter and Downtime Count metrics. Alert when a server transitions state more than 2-3 times within a 15-minute window.
Recommended action:
Review health check configuration - increase check intervals or adjust timeouts if checks are too aggressive. Investigate network path between HAProxy and backends for packet loss or latency spikes. Consider using fall/rise parameters to require multiple consecutive failures before marking servers down.