mirror of
https://gitee.com/wanwujie/sub2api
synced 2026-04-26 01:14:47 +08:00
Reduce ops and scheduler write amplification
This commit is contained in:
ius
@@ -121,14 +121,74 @@ func (s *OpsService) IsMonitoringEnabled(ctx context.Context) bool {
|
||||
}
|
||||
|
||||
func (s *OpsService) RecordError(ctx context.Context, entry *OpsInsertErrorLogInput, rawRequestBody []byte) error {
|
||||
if entry == nil {
|
||||
prepared, ok, err := s.prepareErrorLogInput(ctx, entry, rawRequestBody)
|
||||
if err != nil {
|
||||
log.Printf("[Ops] RecordError prepare failed: %v", err)
|
||||
return err
|
||||
}
|
||||
if !ok {
|
||||
return nil
|
||||
}
|
||||
|
||||
if _, err := s.opsRepo.InsertErrorLog(ctx, prepared); err != nil {
|
||||
// Never bubble up to gateway; best-effort logging.
|
||||
log.Printf("[Ops] RecordError failed: %v", err)
|
||||
return err
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
func (s *OpsService) RecordErrorBatch(ctx context.Context, entries []*OpsInsertErrorLogInput) error {
|
||||
if len(entries) == 0 {
|
||||
return nil
|
||||
}
|
||||
prepared := make([]*OpsInsertErrorLogInput, 0, len(entries))
|
||||
for _, entry := range entries {
|
||||
item, ok, err := s.prepareErrorLogInput(ctx, entry, nil)
|
||||
if err != nil {
|
||||
log.Printf("[Ops] RecordErrorBatch prepare failed: %v", err)
|
||||
continue
|
||||
}
|
||||
if ok {
|
||||
prepared = append(prepared, item)
|
||||
}
|
||||
}
|
||||
if len(prepared) == 0 {
|
||||
return nil
|
||||
}
|
||||
if len(prepared) == 1 {
|
||||
_, err := s.opsRepo.InsertErrorLog(ctx, prepared[0])
|
||||
if err != nil {
|
||||
log.Printf("[Ops] RecordErrorBatch single insert failed: %v", err)
|
||||
}
|
||||
return err
|
||||
}
|
||||
|
||||
if _, err := s.opsRepo.BatchInsertErrorLogs(ctx, prepared); err != nil {
|
||||
log.Printf("[Ops] RecordErrorBatch failed, fallback to single inserts: %v", err)
|
||||
var firstErr error
|
||||
for _, entry := range prepared {
|
||||
if _, insertErr := s.opsRepo.InsertErrorLog(ctx, entry); insertErr != nil {
|
||||
log.Printf("[Ops] RecordErrorBatch fallback insert failed: %v", insertErr)
|
||||
if firstErr == nil {
|
||||
firstErr = insertErr
|
||||
}
|
||||
}
|
||||
}
|
||||
return firstErr
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
func (s *OpsService) prepareErrorLogInput(ctx context.Context, entry *OpsInsertErrorLogInput, rawRequestBody []byte) (*OpsInsertErrorLogInput, bool, error) {
|
||||
if entry == nil {
|
||||
return nil, false, nil
|
||||
}
|
||||
if !s.IsMonitoringEnabled(ctx) {
|
||||
return nil
|
||||
return nil, false, nil
|
||||
}
|
||||
if s.opsRepo == nil {
|
||||
return nil
|
||||
return nil, false, nil
|
||||
}
|
||||
|
||||
// Ensure timestamps are always populated.
|
||||
@@ -185,85 +245,88 @@ func (s *OpsService) RecordError(ctx context.Context, entry *OpsInsertErrorLogIn
|
||||
}
|
||||
}
|
||||
|
||||
// Sanitize + serialize upstream error events list.
|
||||
if len(entry.UpstreamErrors) > 0 {
|
||||
const maxEvents = 32
|
||||
events := entry.UpstreamErrors
|
||||
if len(events) > maxEvents {
|
||||
events = events[len(events)-maxEvents:]
|
||||
if err := sanitizeOpsUpstreamErrors(entry); err != nil {
|
||||
return nil, false, err
|
||||
}
|
||||
|
||||
return entry, true, nil
|
||||
}
|
||||
|
||||
func sanitizeOpsUpstreamErrors(entry *OpsInsertErrorLogInput) error {
|
||||
if entry == nil || len(entry.UpstreamErrors) == 0 {
|
||||
return nil
|
||||
}
|
||||
|
||||
const maxEvents = 32
|
||||
events := entry.UpstreamErrors
|
||||
if len(events) > maxEvents {
|
||||
events = events[len(events)-maxEvents:]
|
||||
}
|
||||
|
||||
sanitized := make([]*OpsUpstreamErrorEvent, 0, len(events))
|
||||
for _, ev := range events {
|
||||
if ev == nil {
|
||||
continue
|
||||
}
|
||||
out := *ev
|
||||
|
||||
out.Platform = strings.TrimSpace(out.Platform)
|
||||
out.UpstreamRequestID = truncateString(strings.TrimSpace(out.UpstreamRequestID), 128)
|
||||
out.Kind = truncateString(strings.TrimSpace(out.Kind), 64)
|
||||
|
||||
if out.AccountID < 0 {
|
||||
out.AccountID = 0
|
||||
}
|
||||
if out.UpstreamStatusCode < 0 {
|
||||
out.UpstreamStatusCode = 0
|
||||
}
|
||||
if out.AtUnixMs < 0 {
|
||||
out.AtUnixMs = 0
|
||||
}
|
||||
|
||||
sanitized := make([]*OpsUpstreamErrorEvent, 0, len(events))
|
||||
for _, ev := range events {
|
||||
if ev == nil {
|
||||
continue
|
||||
}
|
||||
out := *ev
|
||||
msg := sanitizeUpstreamErrorMessage(strings.TrimSpace(out.Message))
|
||||
msg = truncateString(msg, 2048)
|
||||
out.Message = msg
|
||||
|
||||
out.Platform = strings.TrimSpace(out.Platform)
|
||||
out.UpstreamRequestID = truncateString(strings.TrimSpace(out.UpstreamRequestID), 128)
|
||||
out.Kind = truncateString(strings.TrimSpace(out.Kind), 64)
|
||||
detail := strings.TrimSpace(out.Detail)
|
||||
if detail != "" {
|
||||
// Keep upstream detail small; request bodies are not stored here, only upstream error payloads.
|
||||
sanitizedDetail, _ := sanitizeErrorBodyForStorage(detail, opsMaxStoredErrorBodyBytes)
|
||||
out.Detail = sanitizedDetail
|
||||
} else {
|
||||
out.Detail = ""
|
||||
}
|
||||
|
||||
if out.AccountID < 0 {
|
||||
out.AccountID = 0
|
||||
}
|
||||
if out.UpstreamStatusCode < 0 {
|
||||
out.UpstreamStatusCode = 0
|
||||
}
|
||||
if out.AtUnixMs < 0 {
|
||||
out.AtUnixMs = 0
|
||||
}
|
||||
|
||||
msg := sanitizeUpstreamErrorMessage(strings.TrimSpace(out.Message))
|
||||
msg = truncateString(msg, 2048)
|
||||
out.Message = msg
|
||||
|
||||
detail := strings.TrimSpace(out.Detail)
|
||||
if detail != "" {
|
||||
// Keep upstream detail small; request bodies are not stored here, only upstream error payloads.
|
||||
sanitizedDetail, _ := sanitizeErrorBodyForStorage(detail, opsMaxStoredErrorBodyBytes)
|
||||
out.Detail = sanitizedDetail
|
||||
} else {
|
||||
out.Detail = ""
|
||||
}
|
||||
|
||||
out.UpstreamRequestBody = strings.TrimSpace(out.UpstreamRequestBody)
|
||||
if out.UpstreamRequestBody != "" {
|
||||
// Reuse the same sanitization/trimming strategy as request body storage.
|
||||
// Keep it small so it is safe to persist in ops_error_logs JSON.
|
||||
sanitized, truncated, _ := sanitizeAndTrimRequestBody([]byte(out.UpstreamRequestBody), 10*1024)
|
||||
if sanitized != "" {
|
||||
out.UpstreamRequestBody = sanitized
|
||||
if truncated {
|
||||
out.Kind = strings.TrimSpace(out.Kind)
|
||||
if out.Kind == "" {
|
||||
out.Kind = "upstream"
|
||||
}
|
||||
out.Kind = out.Kind + ":request_body_truncated"
|
||||
out.UpstreamRequestBody = strings.TrimSpace(out.UpstreamRequestBody)
|
||||
if out.UpstreamRequestBody != "" {
|
||||
// Reuse the same sanitization/trimming strategy as request body storage.
|
||||
// Keep it small so it is safe to persist in ops_error_logs JSON.
|
||||
sanitizedBody, truncated, _ := sanitizeAndTrimRequestBody([]byte(out.UpstreamRequestBody), 10*1024)
|
||||
if sanitizedBody != "" {
|
||||
out.UpstreamRequestBody = sanitizedBody
|
||||
if truncated {
|
||||
out.Kind = strings.TrimSpace(out.Kind)
|
||||
if out.Kind == "" {
|
||||
out.Kind = "upstream"
|
||||
}
|
||||
} else {
|
||||
out.UpstreamRequestBody = ""
|
||||
out.Kind = out.Kind + ":request_body_truncated"
|
||||
}
|
||||
} else {
|
||||
out.UpstreamRequestBody = ""
|
||||
}
|
||||
|
||||
// Drop fully-empty events (can happen if only status code was known).
|
||||
if out.UpstreamStatusCode == 0 && out.Message == "" && out.Detail == "" {
|
||||
continue
|
||||
}
|
||||
|
||||
evCopy := out
|
||||
sanitized = append(sanitized, &evCopy)
|
||||
}
|
||||
|
||||
entry.UpstreamErrorsJSON = marshalOpsUpstreamErrors(sanitized)
|
||||
entry.UpstreamErrors = nil
|
||||
// Drop fully-empty events (can happen if only status code was known).
|
||||
if out.UpstreamStatusCode == 0 && out.Message == "" && out.Detail == "" {
|
||||
continue
|
||||
}
|
||||
|
||||
evCopy := out
|
||||
sanitized = append(sanitized, &evCopy)
|
||||
}
|
||||
|
||||
if _, err := s.opsRepo.InsertErrorLog(ctx, entry); err != nil {
|
||||
// Never bubble up to gateway; best-effort logging.
|
||||
log.Printf("[Ops] RecordError failed: %v", err)
|
||||
return err
|
||||
}
|
||||
entry.UpstreamErrorsJSON = marshalOpsUpstreamErrors(sanitized)
|
||||
entry.UpstreamErrors = nil
|
||||
return nil
|
||||
}
|
||||
|
||||
|
||||
Reference in New Issue
Block a user