wanwu/sub2api
1
0
Fork 0
mirror of https://gitee.com/wanwujie/sub2api synced 2026-04-26 09:24:48 +08:00
Code Issues Packages Projects Releases Wiki Activity

feat: optimize MODEL_CAPACITY_EXHAUSTED retry and remove extra failover retries

Browse Source 
- MODEL_CAPACITY_EXHAUSTED now uses independent retry strategy:
  - retryDelay < 20s: wait actual retryDelay then retry once
  - retryDelay >= 20s or missing: retry up to 5 times at 20s intervals
  - Still capacity exhausted after retries: switch account (failover)
  - Different error during retry (e.g. 429): handle by actual error code
  - No model rate limit set (capacity != rate limit)

- Remove Antigravity extra failover retries feature:
  Same-account retry mechanism (cherry-picked) makes it redundant.
  Removed: antigravityExtraRetries config, sleepFixedDelay, skip-non-antigravity logic.
This commit is contained in:
erio
2026-02-10 03:47:40 +08:00
parent 6328e69441
commit 662625a091
7 changed files with 282 additions and 559 deletions
Download Patch File Download Diff File Expand all files Collapse all files

65
backend/internal/handler/gateway_handler.go
View File

@@ -39,7 +39,6 @@ type GatewayHandler struct {
concurrencyHelper *ConcurrencyHelper
maxAccountSwitches int
maxAccountSwitchesGemini int
antigravityExtraRetries int
}
// NewGatewayHandler creates a new GatewayHandler
@@ -58,7 +57,6 @@ func NewGatewayHandler(
pingInterval := time.Duration(0)
maxAccountSwitches := 10
maxAccountSwitchesGemini := 3
antigravityExtraRetries := 10
if cfg != nil {
pingInterval = time.Duration(cfg.Concurrency.PingInterval) * time.Second
if cfg.Gateway.MaxAccountSwitches > 0 {
@@ -67,7 +65,6 @@ func NewGatewayHandler(
if cfg.Gateway.MaxAccountSwitchesGemini > 0 {
maxAccountSwitchesGemini = cfg.Gateway.MaxAccountSwitchesGemini
}
antigravityExtraRetries = cfg.Gateway.AntigravityExtraRetries
}
return &GatewayHandler{
gatewayService: gatewayService,
@@ -81,7 +78,6 @@ func NewGatewayHandler(
concurrencyHelper: NewConcurrencyHelper(concurrencyService, SSEPingFormatClaude, pingInterval),
maxAccountSwitches: maxAccountSwitches,
maxAccountSwitchesGemini: maxAccountSwitchesGemini,
antigravityExtraRetries: antigravityExtraRetries,
}
}
@@ -238,7 +234,6 @@ func (h *GatewayHandler) Messages(c *gin.Context) {
if platform == service.PlatformGemini {
maxAccountSwitches := h.maxAccountSwitchesGemini
switchCount := 0
antigravityExtraCount := 0
failedAccountIDs := make(map[int64]struct{})
sameAccountRetryCount := make(map[int64]int) // 同账号重试计数
var lastFailoverErr *service.UpstreamFailoverError
@@ -261,15 +256,6 @@ func (h *GatewayHandler) Messages(c *gin.Context) {
account := selection.Account
setOpsSelectedAccount(c, account.ID)
// 额外重试阶段：跳过非 Antigravity 账号
if switchCount >= maxAccountSwitches && account.Platform != service.PlatformAntigravity {
failedAccountIDs[account.ID] = struct{}{}
if selection.Acquired && selection.ReleaseFunc != nil {
selection.ReleaseFunc()
}
continue
}
// 检查请求拦截（预热请求、SUGGESTION MODE等）
if account.IsInterceptWarmupEnabled() {
interceptType := detectInterceptType(body, reqModel, parsedReq.MaxTokens, reqStream, isClaudeCodeClient)
@@ -377,17 +363,8 @@ func (h *GatewayHandler) Messages(c *gin.Context) {
failedAccountIDs[account.ID] = struct{}{}
if switchCount >= maxAccountSwitches {
// 默认重试用完，进入 Antigravity 额外重试
antigravityExtraCount++
if antigravityExtraCount > h.antigravityExtraRetries {
h.handleFailoverExhausted(c, failoverErr, service.PlatformGemini, streamStarted)
return
}
log.Printf("Account %d: antigravity extra retry %d/%d", account.ID, antigravityExtraCount, h.antigravityExtraRetries)
if !sleepFixedDelay(c.Request.Context(), antigravityExtraRetryDelay) {
return
}
continue
h.handleFailoverExhausted(c, failoverErr, service.PlatformGemini, streamStarted)
return
}
switchCount++
log.Printf("Account %d: upstream error %d, switching account %d/%d", account.ID, failoverErr.StatusCode, switchCount, maxAccountSwitches)
@@ -440,7 +417,6 @@ func (h *GatewayHandler) Messages(c *gin.Context) {
for {
maxAccountSwitches := h.maxAccountSwitches
switchCount := 0
antigravityExtraCount := 0
failedAccountIDs := make(map[int64]struct{})
sameAccountRetryCount := make(map[int64]int) // 同账号重试计数
var lastFailoverErr *service.UpstreamFailoverError
@@ -465,15 +441,6 @@ func (h *GatewayHandler) Messages(c *gin.Context) {
account := selection.Account
setOpsSelectedAccount(c, account.ID)
// 额外重试阶段：跳过非 Antigravity 账号
if switchCount >= maxAccountSwitches && account.Platform != service.PlatformAntigravity {
failedAccountIDs[account.ID] = struct{}{}
if selection.Acquired && selection.ReleaseFunc != nil {
selection.ReleaseFunc()
}
continue
}
// 检查请求拦截（预热请求、SUGGESTION MODE等）
if account.IsInterceptWarmupEnabled() {
interceptType := detectInterceptType(body, reqModel, parsedReq.MaxTokens, reqStream, isClaudeCodeClient)
@@ -614,17 +581,8 @@ func (h *GatewayHandler) Messages(c *gin.Context) {
failedAccountIDs[account.ID] = struct{}{}
if switchCount >= maxAccountSwitches {
// 默认重试用完，进入 Antigravity 额外重试
antigravityExtraCount++
if antigravityExtraCount > h.antigravityExtraRetries {
h.handleFailoverExhausted(c, failoverErr, account.Platform, streamStarted)
return
}
log.Printf("Account %d: antigravity extra retry %d/%d", account.ID, antigravityExtraCount, h.antigravityExtraRetries)
if !sleepFixedDelay(c.Request.Context(), antigravityExtraRetryDelay) {
return
}
continue
h.handleFailoverExhausted(c, failoverErr, account.Platform, streamStarted)
return
}
switchCount++
log.Printf("Account %d: upstream error %d, switching account %d/%d", account.ID, failoverErr.StatusCode, switchCount, maxAccountSwitches)
@@ -933,21 +891,6 @@ func sleepFailoverDelay(ctx context.Context, switchCount int) bool {
}
}
const antigravityExtraRetryDelay = 500 * time.Millisecond
// sleepFixedDelay 固定延时等待，返回 false 表示 context 已取消。
func sleepFixedDelay(ctx context.Context, delay time.Duration) bool {
if delay <= 0 {
return true
}
select {
case <-ctx.Done():
return false
case <-time.After(delay):
return true
}
}
func (h *GatewayHandler) handleFailoverExhausted(c *gin.Context, failoverErr *service.UpstreamFailoverError, platform string, streamStarted bool) {
statusCode := failoverErr.StatusCode
responseBody := failoverErr.ResponseBody