This commit captures both the prior accumulated work-in-progress
(framework migration web/→svelte/, postgres storage, conversation
viewer, dashboard auth, OpenAPI spec, integration tests) AND today's
operational improvements layered on top. History wasn't checkpointed
incrementally; happy to split it via interactive rebase if a reviewer
wants smaller commits.
Today's changes (in addition to the older WIP):
1. Configurable upstream response-header timeout
- ANTHROPIC_RESPONSE_HEADER_TIMEOUT env (default 300s)
- Replaces hardcoded 300s in provider/anthropic.go that was firing
on opus + 1M-context + extended thinking non-streaming requests
- Files: internal/config/config.go, internal/provider/anthropic.go
2. Structured forward-error diagnostic logging
- When a forward to Anthropic fails, log a single key=value line
with request_id, model, stream, body_bytes, has_thinking,
anthropic_beta, query, elapsed, ctx_err — alongside the existing
human-readable error line for back-compat
- Files: internal/handler/handlers.go (logForwardFailure)
3. Full SSE protocol passthrough + Flusher fix
- handler/handlers.go: forward all SSE lines verbatim (event:, id:,
retry:, : comments, blank-line terminators), not only data:.
Previous code produced malformed SSE for strict parsers.
- middleware/logging.go: explicit Flush() method on responseWriter.
Embedding http.ResponseWriter (interface) does not auto-promote
Flush(), so every w.(http.Flusher) check in the streaming
handler was returning ok=false and SSE writes buffered in net/http
until the body closed.
4. Non-streaming → streaming demotion (feature-flagged)
- ANTHROPIC_DEMOTE_NONSTREAMING env (default false)
- When enabled and the routed provider is anthropic, force stream=true
upstream for clients that asked for stream=false. Receive SSE,
accumulate via accumulateSSEToMessage (handles text, tool_use with
partial_json reassembly, thinking, signature, citations_delta,
usage merge), and synthesize a single non-streaming JSON response.
- Eliminates the ResponseHeaderTimeout class of failure entirely.
- Body rewrite uses json.Decoder + UseNumber() to preserve integer
precision in unknown nested fields (tool inputs from prior turns).
- Files: internal/config/config.go, internal/handler/handlers.go,
cmd/proxy/main.go, cmd/proxy/main_test.go
5. Live operational state: /livez gauge + graceful drain
- New internal/runtime package: atomic in-flight counter + draining flag
- New middleware/inflight.go: increments runtime gauge, applied to
/v1/* subrouter so Messages, ChatCompletions, and ProxyPassthrough
are all counted
- /v1/* moved to a gorilla/mux subrouter so the InFlight middleware
applies surgically; /health, /livez, /openapi.* remain on parent
router (unauthenticated, uncounted)
- Health handler returns 503 draining when runtime.IsDraining() is
true, so Traefik stops routing to a slot before drain begins
- New /livez handler returns {status, in_flight, draining, timestamp}
- SIGTERM handler in main.go: SetDraining(true), poll for in_flight==0
with 32-min ceiling and 1s tick (logs every 10s), then srv.Shutdown
- Auth bypass list extended with /livez
- Files: internal/runtime/runtime.go (new),
internal/middleware/inflight.go (new),
internal/middleware/auth.go,
internal/handler/handlers.go (Health, Livez, runtime import),
cmd/proxy/main.go (subrouter, drain loop)
6. OpenAPI spec updates
- Document Health 503 response and new DrainingResponse schema
- Add /livez path with LivezResponse schema
- Files: internal/handler/openapi.go
Verified: go build ./... clean, go test ./... all pass, go vet clean.
Three rounds of codex peer review across changes 1-5; all feedback
addressed (citations_delta, json.Number precision, drain-loop logging
via lastLog timestamp, PathPrefix tightened to "/v1/").
203 lines
5.1 KiB
Go
203 lines
5.1 KiB
Go
package service
|
|
|
|
import (
|
|
"encoding/json"
|
|
"log"
|
|
"strings"
|
|
|
|
"github.com/seifghazi/claude-code-monitor/internal/config"
|
|
"github.com/seifghazi/claude-code-monitor/internal/model"
|
|
)
|
|
|
|
const redactionPlaceholder = "[REDACTED]"
|
|
|
|
// maxStoredBodyBytes is the maximum serialized size of a request body stored in the DB.
|
|
// Bodies larger than this (e.g. 1M context payloads) are replaced with a metadata summary.
|
|
const maxStoredBodyBytes = 512 * 1024 // 512 KB
|
|
|
|
func prepareRequestBodyForStorage(cfg *config.StorageConfig, body interface{}) (interface{}, error) {
|
|
if shouldSuppressBodies(cfg) {
|
|
return storageBodyPlaceholder("metadata_only"), nil
|
|
}
|
|
if cfg != nil && !cfg.CaptureRequestBody {
|
|
return storageBodyPlaceholder("request_body_disabled"), nil
|
|
}
|
|
|
|
normalized, err := normalizeJSONValue(body)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
redacted := redactJSONValue(normalized, redactedFieldSet(redactedFields(cfg)))
|
|
|
|
// Check serialized size; if too large, store a lightweight summary instead
|
|
data, err := json.Marshal(redacted)
|
|
if err != nil {
|
|
return redacted, nil
|
|
}
|
|
if len(data) > maxStoredBodyBytes {
|
|
return truncatedBodySummary(redacted, len(data)), nil
|
|
}
|
|
|
|
return redacted, nil
|
|
}
|
|
|
|
// truncatedBodySummary extracts key metadata from an oversized request body
|
|
// so the DB row stays small while retaining useful diagnostic info.
|
|
func truncatedBodySummary(body interface{}, originalBytes int) map[string]interface{} {
|
|
summary := map[string]interface{}{
|
|
"_truncated": true,
|
|
"_original_bytes": originalBytes,
|
|
}
|
|
if m, ok := body.(map[string]interface{}); ok {
|
|
if v, ok := m["model"]; ok {
|
|
summary["model"] = v
|
|
}
|
|
if v, ok := m["stream"]; ok {
|
|
summary["stream"] = v
|
|
}
|
|
if v, ok := m["max_tokens"]; ok {
|
|
summary["max_tokens"] = v
|
|
}
|
|
if msgs, ok := m["messages"].([]interface{}); ok {
|
|
summary["message_count"] = len(msgs)
|
|
}
|
|
if sys, ok := m["system"].([]interface{}); ok {
|
|
summary["system_count"] = len(sys)
|
|
}
|
|
if tools, ok := m["tools"].([]interface{}); ok {
|
|
summary["tool_count"] = len(tools)
|
|
}
|
|
}
|
|
return summary
|
|
}
|
|
|
|
func prepareResponseForStorage(cfg *config.StorageConfig, logger *log.Logger, response *model.ResponseLog) (*model.ResponseLog, error) {
|
|
if response == nil {
|
|
return nil, nil
|
|
}
|
|
|
|
clone := *response
|
|
if shouldSuppressBodies(cfg) || (cfg != nil && !cfg.CaptureResponseBody) {
|
|
clone.Body = nil
|
|
clone.BodyText = ""
|
|
clone.StreamingChunks = nil
|
|
clone.ChunkTimings = nil
|
|
return &clone, nil
|
|
}
|
|
|
|
if len(clone.Body) > 0 {
|
|
sanitizedBody, err := sanitizeRawJSON(clone.Body, redactedFieldSet(redactedFields(cfg)))
|
|
if err != nil {
|
|
if logger != nil {
|
|
logger.Printf("Warning: failed to redact response body: %v", err)
|
|
}
|
|
} else {
|
|
clone.Body = sanitizedBody
|
|
}
|
|
// Cap stored response body size
|
|
if len(clone.Body) > maxStoredBodyBytes {
|
|
clone.Body = json.RawMessage(`{"_truncated":true}`)
|
|
}
|
|
}
|
|
|
|
// Cap stored streaming chunks to avoid huge DB rows on long streams
|
|
const maxStoredChunks = 500
|
|
if len(clone.StreamingChunks) > maxStoredChunks {
|
|
clone.StreamingChunks = clone.StreamingChunks[:maxStoredChunks]
|
|
}
|
|
if len(clone.ChunkTimings) > maxStoredChunks {
|
|
clone.ChunkTimings = clone.ChunkTimings[:maxStoredChunks]
|
|
}
|
|
|
|
return &clone, nil
|
|
}
|
|
|
|
func shouldSuppressBodies(cfg *config.StorageConfig) bool {
|
|
return cfg != nil && cfg.MetadataOnly
|
|
}
|
|
|
|
func redactedFields(cfg *config.StorageConfig) []string {
|
|
if cfg == nil {
|
|
return nil
|
|
}
|
|
return cfg.RedactedFields
|
|
}
|
|
|
|
func normalizeJSONValue(value interface{}) (interface{}, error) {
|
|
if value == nil {
|
|
return nil, nil
|
|
}
|
|
|
|
data, err := json.Marshal(value)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
var normalized interface{}
|
|
if err := json.Unmarshal(data, &normalized); err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
return normalized, nil
|
|
}
|
|
|
|
func sanitizeRawJSON(raw json.RawMessage, redacted map[string]struct{}) (json.RawMessage, error) {
|
|
if len(raw) == 0 {
|
|
return raw, nil
|
|
}
|
|
|
|
var value interface{}
|
|
if err := json.Unmarshal(raw, &value); err != nil {
|
|
return raw, err
|
|
}
|
|
|
|
sanitized := redactJSONValue(value, redacted)
|
|
data, err := json.Marshal(sanitized)
|
|
if err != nil {
|
|
return raw, err
|
|
}
|
|
|
|
return json.RawMessage(data), nil
|
|
}
|
|
|
|
func redactJSONValue(value interface{}, redacted map[string]struct{}) interface{} {
|
|
switch typed := value.(type) {
|
|
case map[string]interface{}:
|
|
result := make(map[string]interface{}, len(typed))
|
|
for key, child := range typed {
|
|
if _, ok := redacted[strings.ToLower(key)]; ok {
|
|
result[key] = redactionPlaceholder
|
|
continue
|
|
}
|
|
result[key] = redactJSONValue(child, redacted)
|
|
}
|
|
return result
|
|
case []interface{}:
|
|
result := make([]interface{}, len(typed))
|
|
for i, child := range typed {
|
|
result[i] = redactJSONValue(child, redacted)
|
|
}
|
|
return result
|
|
default:
|
|
return value
|
|
}
|
|
}
|
|
|
|
func storageBodyPlaceholder(mode string) map[string]interface{} {
|
|
return map[string]interface{}{
|
|
"_storage_mode": mode,
|
|
}
|
|
}
|
|
|
|
func redactedFieldSet(fields []string) map[string]struct{} {
|
|
set := make(map[string]struct{}, len(fields))
|
|
for _, field := range fields {
|
|
field = strings.TrimSpace(strings.ToLower(field))
|
|
if field == "" {
|
|
continue
|
|
}
|
|
set[field] = struct{}{}
|
|
}
|
|
return set
|
|
}
|