lukaszraczylo/traefikoidc
1
0
Fork 0
mirror of https://github.com/lukaszraczylo/traefikoidc.git synced 2026-11-09 08:55:36 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix(resilience): recover from unhealthy and close half-open on successes

Browse Source 
The health check promoted to healthy only when the fail counter was
above zero. The first success zeroed that counter, so a threshold of 2
or more never recovered. It now promotes on the status and the success
streak.

The breaker closed when admitted probes reached HalfOpenMaxRequests,
not when that many succeeded. A new counter tracks successes.
This commit is contained in:
lukaszraczylo
2026-11-03 00:46:46 +00:00
parent ac0fcfbe6a
commit eec076f812
3 changed files with 66 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files
internal/cache/resilience/health_check_backend.go
Vendored
+4 -3
View File
@@ -168,11 +168,12 @@ func (h *HealthCheckBackend) IsHealthy() bool {
func (h *HealthCheckBackend) recordResult(success bool) {
// #nosec G115 -- threshold config values are small integers that fit in int32
if success {
fails := h.consecutiveFails.Swap(0)
h.consecutiveFails.Store(0)
oks := h.consecutiveOK.Add(1)
// Check if we should transition to healthy
if fails > 0 && oks >= int32(h.config.HealthyThreshold) {
// Promote on a success streak. Keyed on the status, not on the fail
// counter, which the first success of the streak already zeroed.
if HealthStatus(h.status.Load()) != HealthHealthy && oks >= int32(h.config.HealthyThreshold) {
oldStatus := HealthStatus(h.status.Swap(int32(HealthHealthy)))
if oldStatus != HealthHealthy && h.config.OnStatusChange != nil {
h.config.OnStatusChange(oldStatus, HealthHealthy)