Reverse Proxy 502 — All Backend Servers Health Check Failed
HAProxy is returning 502 Bad Gateway because all backend servers failed their health checks simultaneously after a coordinated deployment.
Pattern
CONNECTION_REFUSED
Expected: PROXY_FAILURE
Severity
CRITICAL
Confidence
68%
Remediation
Auto-Heal
Test Results
Metric
Expected
Actual
Result
Pattern Recognition
PROXY_FAILURE
CONNECTION_REFUSED
Severity Assessment
CRITICAL
CRITICAL
Incident Correlation
N/A
None
Cascade Escalation
N/A
No
Remediation
—
Auto-Heal — Corax resolves autonomously
Scenario Conditions
HAProxy 2.8. Backend pool 'api-servers' with 4 servers. All 4 deployed at same time. New version crashes on startup. HAProxy health checks: all DOWN.
Injected Error Messages (1)
HAProxy 502 — all 4 backend servers in pool 'api-servers' DOWN, health check failing on all: connection refused, simultaneous deployment crashed all backends, 100% error rate
Neural Engine Root Cause Analysis
Connection refused — service is likely crashed or not listening on the expected port.
Remediation Plan
Restart the target service. If recurring, check for memory leaks or crash loops.
Improvements Applied
Pattern classified as CONNECTION_REFUSED (expected PROXY_FAILURE)