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

feat(ops): 实现上游错误事件记录与查询功能

Browse Source 
**新增功能**:
- 新建ops_upstream_error_events表存储上游服务错误详情
- 支持记录上游429/529/5xx错误的详细上下文信息
- 提供按时间范围查询上游错误事件的API

**后端改动**:
1. 模型层（ops_models.go, ops_port.go）:
   - 新增UpstreamErrorEvent结构体
   - 扩展Repository接口支持上游错误事件CRUD

2. 仓储层（ops_repo.go）:
   - 实现InsertUpstreamErrorEvent写入上游错误
   - 实现GetUpstreamErrorEvents按时间范围查询

3. 服务层（ops_service.go, ops_upstream_context.go）:
   - ops_service: 新增GetUpstreamErrorEvents查询方法
   - ops_upstream_context: 封装上游错误上下文构建逻辑

4. Handler层（ops_error_logger.go）:
   - 新增GetUpstreamErrorsHandler处理上游错误查询请求

5. Gateway层集成:
   - antigravity_gateway_service.go: 429/529错误时记录上游事件
   - gateway_service.go: OpenAI 429/5xx错误时记录
   - gemini_messages_compat_service.go: Gemini 429/5xx错误时记录
   - openai_gateway_service.go: OpenAI 429/5xx错误时记录
   - ratelimit_service.go: 429限流错误时记录

**数据记录字段**:
- request_id: 关联ops_logs主记录
- platform/model: 上游服务标识
- status_code/error_message: 错误详情
- request_headers/response_body: 调试信息（可选）
- created_at: 错误发生时间
This commit is contained in:
IanShaw027
2026-01-11 15:30:27 +08:00
parent c2962752eb
commit 7ebca553ef
11 changed files with 907 additions and 41 deletions
Download Patch File Download Diff File Expand all files Collapse all files

256
backend/internal/service/antigravity_gateway_service.go
View File

@@ -564,6 +564,14 @@ urlFallbackLoop:
resp, err = s.httpUpstream.Do(upstreamReq, proxyURL, account.ID, account.Concurrency)
if err != nil {
safeErr := sanitizeUpstreamErrorMessage(err.Error())
appendOpsUpstreamError(c, OpsUpstreamErrorEvent{
Platform: account.Platform,
AccountID: account.ID,
UpstreamStatusCode: 0,
Kind: "request_error",
Message: safeErr,
})
// 检查是否应触发 URL 降级
if shouldAntigravityFallbackToNextURL(err, 0) && urlIdx < len(availableURLs)-1 {
antigravity.DefaultURLAvailability.MarkUnavailable(baseURL)
@@ -579,6 +587,7 @@ urlFallbackLoop:
continue
}
log.Printf("%s status=request_failed retries_exhausted error=%v", prefix, err)
setOpsUpstreamError(c, 0, safeErr, "")
return nil, s.writeClaudeError(c, http.StatusBadGateway, "upstream_error", "Upstream request failed after retries")
}
@@ -586,6 +595,26 @@ urlFallbackLoop:
if resp.StatusCode == http.StatusTooManyRequests && urlIdx < len(availableURLs)-1 {
respBody, _ := io.ReadAll(io.LimitReader(resp.Body, 2<<20))
_ = resp.Body.Close()
upstreamMsg := strings.TrimSpace(extractAntigravityErrorMessage(respBody))
upstreamMsg = sanitizeUpstreamErrorMessage(upstreamMsg)
logBody := s.settingService != nil && s.settingService.cfg != nil && s.settingService.cfg.Gateway.LogUpstreamErrorBody
maxBytes := 2048
if s.settingService != nil && s.settingService.cfg != nil && s.settingService.cfg.Gateway.LogUpstreamErrorBodyMaxBytes > 0 {
maxBytes = s.settingService.cfg.Gateway.LogUpstreamErrorBodyMaxBytes
}
upstreamDetail := ""
if logBody {
upstreamDetail = truncateString(string(respBody), maxBytes)
}
appendOpsUpstreamError(c, OpsUpstreamErrorEvent{
Platform: account.Platform,
AccountID: account.ID,
UpstreamStatusCode: resp.StatusCode,
UpstreamRequestID: resp.Header.Get("x-request-id"),
Kind: "retry",
Message: upstreamMsg,
Detail: upstreamDetail,
})
antigravity.DefaultURLAvailability.MarkUnavailable(baseURL)
log.Printf("%s URL fallback (HTTP 429): %s -> %s body=%s", prefix, baseURL, availableURLs[urlIdx+1], truncateForLog(respBody, 200))
continue urlFallbackLoop
@@ -596,6 +625,26 @@ urlFallbackLoop:
_ = resp.Body.Close()
if attempt < antigravityMaxRetries {
upstreamMsg := strings.TrimSpace(extractAntigravityErrorMessage(respBody))
upstreamMsg = sanitizeUpstreamErrorMessage(upstreamMsg)
logBody := s.settingService != nil && s.settingService.cfg != nil && s.settingService.cfg.Gateway.LogUpstreamErrorBody
maxBytes := 2048
if s.settingService != nil && s.settingService.cfg != nil && s.settingService.cfg.Gateway.LogUpstreamErrorBodyMaxBytes > 0 {
maxBytes = s.settingService.cfg.Gateway.LogUpstreamErrorBodyMaxBytes
}
upstreamDetail := ""
if logBody {
upstreamDetail = truncateString(string(respBody), maxBytes)
}
appendOpsUpstreamError(c, OpsUpstreamErrorEvent{
Platform: account.Platform,
AccountID: account.ID,
UpstreamStatusCode: resp.StatusCode,
UpstreamRequestID: resp.Header.Get("x-request-id"),
Kind: "retry",
Message: upstreamMsg,
Detail: upstreamDetail,
})
log.Printf("%s status=%d retry=%d/%d body=%s", prefix, resp.StatusCode, attempt, antigravityMaxRetries, truncateForLog(respBody, 500))
if !sleepAntigravityBackoffWithContext(ctx, attempt) {
log.Printf("%s status=context_canceled_during_backoff", prefix)
@@ -628,6 +677,27 @@ urlFallbackLoop:
// Antigravity /v1internal 链路在部分场景会对 thought/thinking signature 做严格校验，
// 当历史消息携带的 signature 不合法时会直接 400；去除 thinking 后可继续完成请求。
if resp.StatusCode == http.StatusBadRequest && isSignatureRelatedError(respBody) {
upstreamMsg := strings.TrimSpace(extractAntigravityErrorMessage(respBody))
upstreamMsg = sanitizeUpstreamErrorMessage(upstreamMsg)
logBody := s.settingService != nil && s.settingService.cfg != nil && s.settingService.cfg.Gateway.LogUpstreamErrorBody
maxBytes := 2048
if s.settingService != nil && s.settingService.cfg != nil && s.settingService.cfg.Gateway.LogUpstreamErrorBodyMaxBytes > 0 {
maxBytes = s.settingService.cfg.Gateway.LogUpstreamErrorBodyMaxBytes
}
upstreamDetail := ""
if logBody {
upstreamDetail = truncateString(string(respBody), maxBytes)
}
appendOpsUpstreamError(c, OpsUpstreamErrorEvent{
Platform: account.Platform,
AccountID: account.ID,
UpstreamStatusCode: resp.StatusCode,
UpstreamRequestID: resp.Header.Get("x-request-id"),
Kind: "signature_error",
Message: upstreamMsg,
Detail: upstreamDetail,
})
// Conservative two-stage fallback:
// 1) Disable top-level thinking + thinking->text
// 2) Only if still signature-related 400: also downgrade tool_use/tool_result to text.
@@ -661,6 +731,13 @@ urlFallbackLoop:
}
retryResp, retryErr := s.httpUpstream.Do(retryReq, proxyURL, account.ID, account.Concurrency)
if retryErr != nil {
appendOpsUpstreamError(c, OpsUpstreamErrorEvent{
Platform: account.Platform,
AccountID: account.ID,
UpstreamStatusCode: 0,
Kind: "signature_retry_request_error",
Message: sanitizeUpstreamErrorMessage(retryErr.Error()),
})
log.Printf("Antigravity account %d: signature retry request failed (%s): %v", account.ID, stage.name, retryErr)
continue
}
@@ -674,6 +751,25 @@ urlFallbackLoop:
retryBody, _ := io.ReadAll(io.LimitReader(retryResp.Body, 2<<20))
_ = retryResp.Body.Close()
kind := "signature_retry"
if strings.TrimSpace(stage.name) != "" {
kind = "signature_retry_" + strings.ReplaceAll(stage.name, "+", "_")
}
retryUpstreamMsg := strings.TrimSpace(extractAntigravityErrorMessage(retryBody))
retryUpstreamMsg = sanitizeUpstreamErrorMessage(retryUpstreamMsg)
retryUpstreamDetail := ""
if logBody {
retryUpstreamDetail = truncateString(string(retryBody), maxBytes)
}
appendOpsUpstreamError(c, OpsUpstreamErrorEvent{
Platform: account.Platform,
AccountID: account.ID,
UpstreamStatusCode: retryResp.StatusCode,
UpstreamRequestID: retryResp.Header.Get("x-request-id"),
Kind: kind,
Message: retryUpstreamMsg,
Detail: retryUpstreamDetail,
})
// If this stage fixed the signature issue, we stop; otherwise we may try the next stage.
if retryResp.StatusCode != http.StatusBadRequest || !isSignatureRelatedError(retryBody) {
@@ -701,10 +797,30 @@ urlFallbackLoop:
s.handleUpstreamError(ctx, prefix, account, resp.StatusCode, resp.Header, respBody, quotaScope)
if s.shouldFailoverUpstreamError(resp.StatusCode) {
upstreamMsg := strings.TrimSpace(extractAntigravityErrorMessage(respBody))
upstreamMsg = sanitizeUpstreamErrorMessage(upstreamMsg)
logBody := s.settingService != nil && s.settingService.cfg != nil && s.settingService.cfg.Gateway.LogUpstreamErrorBody
maxBytes := 2048
if s.settingService != nil && s.settingService.cfg != nil && s.settingService.cfg.Gateway.LogUpstreamErrorBodyMaxBytes > 0 {
maxBytes = s.settingService.cfg.Gateway.LogUpstreamErrorBodyMaxBytes
}
upstreamDetail := ""
if logBody {
upstreamDetail = truncateString(string(respBody), maxBytes)
}
appendOpsUpstreamError(c, OpsUpstreamErrorEvent{
Platform: account.Platform,
AccountID: account.ID,
UpstreamStatusCode: resp.StatusCode,
UpstreamRequestID: resp.Header.Get("x-request-id"),
Kind: "failover",
Message: upstreamMsg,
Detail: upstreamDetail,
})
return nil, &UpstreamFailoverError{StatusCode: resp.StatusCode}
}
return nil, s.writeMappedClaudeError(c, resp.StatusCode, respBody)
return nil, s.writeMappedClaudeError(c, account, resp.StatusCode, resp.Header.Get("x-request-id"), respBody)
}
}
@@ -1108,6 +1224,14 @@ urlFallbackLoop:
resp, err = s.httpUpstream.Do(upstreamReq, proxyURL, account.ID, account.Concurrency)
if err != nil {
safeErr := sanitizeUpstreamErrorMessage(err.Error())
appendOpsUpstreamError(c, OpsUpstreamErrorEvent{
Platform: account.Platform,
AccountID: account.ID,
UpstreamStatusCode: 0,
Kind: "request_error",
Message: safeErr,
})
// 检查是否应触发 URL 降级
if shouldAntigravityFallbackToNextURL(err, 0) && urlIdx < len(availableURLs)-1 {
antigravity.DefaultURLAvailability.MarkUnavailable(baseURL)
@@ -1123,6 +1247,7 @@ urlFallbackLoop:
continue
}
log.Printf("%s status=request_failed retries_exhausted error=%v", prefix, err)
setOpsUpstreamError(c, 0, safeErr, "")
return nil, s.writeGoogleError(c, http.StatusBadGateway, "Upstream request failed after retries")
}
@@ -1130,6 +1255,26 @@ urlFallbackLoop:
if resp.StatusCode == http.StatusTooManyRequests && urlIdx < len(availableURLs)-1 {
respBody, _ := io.ReadAll(io.LimitReader(resp.Body, 2<<20))
_ = resp.Body.Close()
upstreamMsg := strings.TrimSpace(extractAntigravityErrorMessage(respBody))
upstreamMsg = sanitizeUpstreamErrorMessage(upstreamMsg)
logBody := s.settingService != nil && s.settingService.cfg != nil && s.settingService.cfg.Gateway.LogUpstreamErrorBody
maxBytes := 2048
if s.settingService != nil && s.settingService.cfg != nil && s.settingService.cfg.Gateway.LogUpstreamErrorBodyMaxBytes > 0 {
maxBytes = s.settingService.cfg.Gateway.LogUpstreamErrorBodyMaxBytes
}
upstreamDetail := ""
if logBody {
upstreamDetail = truncateString(string(respBody), maxBytes)
}
appendOpsUpstreamError(c, OpsUpstreamErrorEvent{
Platform: account.Platform,
AccountID: account.ID,
UpstreamStatusCode: resp.StatusCode,
UpstreamRequestID: resp.Header.Get("x-request-id"),
Kind: "retry",
Message: upstreamMsg,
Detail: upstreamDetail,
})
antigravity.DefaultURLAvailability.MarkUnavailable(baseURL)
log.Printf("%s URL fallback (HTTP 429): %s -> %s body=%s", prefix, baseURL, availableURLs[urlIdx+1], truncateForLog(respBody, 200))
continue urlFallbackLoop
@@ -1140,6 +1285,26 @@ urlFallbackLoop:
_ = resp.Body.Close()
if attempt < antigravityMaxRetries {
upstreamMsg := strings.TrimSpace(extractAntigravityErrorMessage(respBody))
upstreamMsg = sanitizeUpstreamErrorMessage(upstreamMsg)
logBody := s.settingService != nil && s.settingService.cfg != nil && s.settingService.cfg.Gateway.LogUpstreamErrorBody
maxBytes := 2048
if s.settingService != nil && s.settingService.cfg != nil && s.settingService.cfg.Gateway.LogUpstreamErrorBodyMaxBytes > 0 {
maxBytes = s.settingService.cfg.Gateway.LogUpstreamErrorBodyMaxBytes
}
upstreamDetail := ""
if logBody {
upstreamDetail = truncateString(string(respBody), maxBytes)
}
appendOpsUpstreamError(c, OpsUpstreamErrorEvent{
Platform: account.Platform,
AccountID: account.ID,
UpstreamStatusCode: resp.StatusCode,
UpstreamRequestID: resp.Header.Get("x-request-id"),
Kind: "retry",
Message: upstreamMsg,
Detail: upstreamDetail,
})
log.Printf("%s status=%d retry=%d/%d", prefix, resp.StatusCode, attempt, antigravityMaxRetries)
if !sleepAntigravityBackoffWithContext(ctx, attempt) {
log.Printf("%s status=context_canceled_during_backoff", prefix)
@@ -1205,21 +1370,59 @@ urlFallbackLoop:
s.handleUpstreamError(ctx, prefix, account, resp.StatusCode, resp.Header, respBody, quotaScope)
if s.shouldFailoverUpstreamError(resp.StatusCode) {
return nil, &UpstreamFailoverError{StatusCode: resp.StatusCode}
}
// 解包并返回错误
requestID := resp.Header.Get("x-request-id")
if requestID != "" {
c.Header("x-request-id", requestID)
}
unwrapped, _ := s.unwrapV1InternalResponse(respBody)
unwrapped, unwrapErr := s.unwrapV1InternalResponse(respBody)
unwrappedForOps := unwrapped
if unwrapErr != nil || len(unwrappedForOps) == 0 {
unwrappedForOps = respBody
}
upstreamMsg := strings.TrimSpace(extractAntigravityErrorMessage(unwrappedForOps))
upstreamMsg = sanitizeUpstreamErrorMessage(upstreamMsg)
logBody := s.settingService != nil && s.settingService.cfg != nil && s.settingService.cfg.Gateway.LogUpstreamErrorBody
maxBytes := 2048
if s.settingService != nil && s.settingService.cfg != nil && s.settingService.cfg.Gateway.LogUpstreamErrorBodyMaxBytes > 0 {
maxBytes = s.settingService.cfg.Gateway.LogUpstreamErrorBodyMaxBytes
}
upstreamDetail := ""
if logBody {
upstreamDetail = truncateString(string(unwrappedForOps), maxBytes)
}
// Always record upstream context for Ops error logs, even when we will failover.
setOpsUpstreamError(c, resp.StatusCode, upstreamMsg, upstreamDetail)
if s.shouldFailoverUpstreamError(resp.StatusCode) {
appendOpsUpstreamError(c, OpsUpstreamErrorEvent{
Platform: account.Platform,
AccountID: account.ID,
UpstreamStatusCode: resp.StatusCode,
UpstreamRequestID: requestID,
Kind: "failover",
Message: upstreamMsg,
Detail: upstreamDetail,
})
return nil, &UpstreamFailoverError{StatusCode: resp.StatusCode}
}
contentType := resp.Header.Get("Content-Type")
if contentType == "" {
contentType = "application/json"
}
c.Data(resp.StatusCode, contentType, unwrapped)
appendOpsUpstreamError(c, OpsUpstreamErrorEvent{
Platform: account.Platform,
AccountID: account.ID,
UpstreamStatusCode: resp.StatusCode,
UpstreamRequestID: requestID,
Kind: "http_error",
Message: upstreamMsg,
Detail: upstreamDetail,
})
c.Data(resp.StatusCode, contentType, unwrappedForOps)
return nil, fmt.Errorf("antigravity upstream error: %d", resp.StatusCode)
}
@@ -1674,9 +1877,35 @@ func (s *AntigravityGatewayService) writeClaudeError(c *gin.Context, status int,
return fmt.Errorf("%s", message)
}
func (s *AntigravityGatewayService) writeMappedClaudeError(c *gin.Context, upstreamStatus int, body []byte) error {
// 记录上游错误详情便于调试
log.Printf("[antigravity-Forward] upstream_error status=%d body=%s", upstreamStatus, string(body))
func (s *AntigravityGatewayService) writeMappedClaudeError(c *gin.Context, account *Account, upstreamStatus int, upstreamRequestID string, body []byte) error {
upstreamMsg := strings.TrimSpace(extractUpstreamErrorMessage(body))
upstreamMsg = sanitizeUpstreamErrorMessage(upstreamMsg)
logBody := s.settingService != nil && s.settingService.cfg != nil && s.settingService.cfg.Gateway.LogUpstreamErrorBody
maxBytes := 2048
if s.settingService != nil && s.settingService.cfg != nil && s.settingService.cfg.Gateway.LogUpstreamErrorBodyMaxBytes > 0 {
maxBytes = s.settingService.cfg.Gateway.LogUpstreamErrorBodyMaxBytes
}
upstreamDetail := ""
if logBody {
upstreamDetail = truncateString(string(body), maxBytes)
}
setOpsUpstreamError(c, upstreamStatus, upstreamMsg, upstreamDetail)
appendOpsUpstreamError(c, OpsUpstreamErrorEvent{
Platform: account.Platform,
AccountID: account.ID,
UpstreamStatusCode: upstreamStatus,
UpstreamRequestID: upstreamRequestID,
Kind: "http_error",
Message: upstreamMsg,
Detail: upstreamDetail,
})
// 记录上游错误详情便于排障（可选：由配置控制；不回显到客户端）
if logBody {
log.Printf("[antigravity-Forward] upstream_error status=%d body=%s", upstreamStatus, truncateForLog(body, maxBytes))
}
var statusCode int
var errType, errMsg string
@@ -1712,7 +1941,10 @@ func (s *AntigravityGatewayService) writeMappedClaudeError(c *gin.Context, upstr
"type": "error",
"error": gin.H{"type": errType, "message": errMsg},
})
return fmt.Errorf("upstream error: %d", upstreamStatus)
if upstreamMsg == "" {
return fmt.Errorf("upstream error: %d", upstreamStatus)
}
return fmt.Errorf("upstream error: %d message=%s", upstreamStatus, upstreamMsg)
}
func (s *AntigravityGatewayService) writeGoogleError(c *gin.Context, status int, message string) error {