oadmin/sub2api
1
0
Fork 0
Code Issues Pull Requests 1 Actions Packages Projects Releases Wiki Activity

fix: ensure sticky session failover triggers cache billing exemption

Browse Source
This commit is contained in:
erio
2026-02-09 06:57:07 +08:00
parent 681950dadd
commit 72b08f9cc5
2 changed files with 9 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
backend/internal/handler/gateway_handler.go
View File

@@ -341,7 +341,7 @@ func (h *GatewayHandler) Messages(c *gin.Context) {
if errors.As(err, &failoverErr) { if errors.As(err, &failoverErr) {
failedAccountIDs[account.ID] = struct{}{} failedAccountIDs[account.ID] = struct{}{}
lastFailoverErr = failoverErr lastFailoverErr = failoverErr
if failoverErr.ForceCacheBilling { if needForceCacheBilling(hasBoundSession, failoverErr) {
forceCacheBilling = true forceCacheBilling = true
} }
if switchCount >= maxAccountSwitches { if switchCount >= maxAccountSwitches {
@@ -541,7 +541,7 @@ func (h *GatewayHandler) Messages(c *gin.Context) {
if errors.As(err, &failoverErr) { if errors.As(err, &failoverErr) {
failedAccountIDs[account.ID] = struct{}{} failedAccountIDs[account.ID] = struct{}{}
lastFailoverErr = failoverErr lastFailoverErr = failoverErr
if failoverErr.ForceCacheBilling { if needForceCacheBilling(hasBoundSession, failoverErr) {
forceCacheBilling = true forceCacheBilling = true
} }
if switchCount >= maxAccountSwitches { if switchCount >= maxAccountSwitches {
@@ -817,6 +817,12 @@ func (h *GatewayHandler) handleConcurrencyError(c *gin.Context, err error, slotT
fmt.Sprintf("Concurrency limit exceeded for %s, please retry later", slotType), streamStarted) fmt.Sprintf("Concurrency limit exceeded for %s, please retry later", slotType), streamStarted)
} }
// needForceCacheBilling 判断 failover 时是否需要强制缓存计费
// 粘性会话切换账号、或上游明确标记时，将 input_tokens 转为 cache_read 计费
func needForceCacheBilling(hasBoundSession bool, failoverErr *service.UpstreamFailoverError) bool {
return hasBoundSession || (failoverErr != nil && failoverErr.ForceCacheBilling)
}
// sleepFailoverDelay 账号切换线性递增延时：第1次0s、第2次1s、第3次2s… // sleepFailoverDelay 账号切换线性递增延时：第1次0s、第2次1s、第3次2s…
// 返回 false 表示 context 已取消。 // 返回 false 表示 context 已取消。
func sleepFailoverDelay(ctx context.Context, switchCount int) bool { func sleepFailoverDelay(ctx context.Context, switchCount int) bool {

2
backend/internal/handler/gemini_v1beta_handler.go
View File

@@ -422,7 +422,7 @@ func (h *GatewayHandler) GeminiV1BetaModels(c *gin.Context) {
var failoverErr *service.UpstreamFailoverError var failoverErr *service.UpstreamFailoverError
if errors.As(err, &failoverErr) { if errors.As(err, &failoverErr) {
failedAccountIDs[account.ID] = struct{}{} failedAccountIDs[account.ID] = struct{}{}
if failoverErr.ForceCacheBilling { if needForceCacheBilling(hasBoundSession, failoverErr) {
forceCacheBilling = true forceCacheBilling = true
} }
if switchCount >= maxAccountSwitches { if switchCount >= maxAccountSwitches {