Files
go-otel/internal/collector/issue_extract.go
T
developerandClaude Sonnet 5 0d274ff89c Derive deduplicated error issues from traces and logs (Phase 2 backend)
- Add internal/issues.Tracker: fingerprints (kind, service, title) into
  deduplicated Issues with count/firstSeen/lastSeen/affected services and
  bounded sample occurrences, evicting the least-recently-seen issue once
  a cap is reached.
- Add internal/collector/issue_extract.go to pull issues out of ingested
  OTLP data: spans with an error status (preferring exception.type/message
  from an "exception" event, falling back to the span status message), and
  logs at ERROR/FATAL severity (by severity number or text).
- Wire extraction into Collector.ConsumeTraces/ConsumeLogs and expose
  GET /api/issues and GET /api/issues/{id}.
- Fix a pre-existing data race in the collector test's mockStore (unsynced
  access to entries while addToHistory appends asynchronously), found via
  `go test -race` while adding the new issue-tracking tests.
- Update docs/PLAN.md, README.md and agents_example.md for the new API.

The UI "Issues" tab itself lives in the separate frontend project that
builds this server's dist/ bundle; this commit only covers the backend.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
2026-09-17 09:35:39 +00:00

162 lines
4.0 KiB
Go

package collector
import (
"strings"
"git.tornberg.me/go-otel/internal/issues"
"go.opentelemetry.io/collector/pdata/pcommon"
"go.opentelemetry.io/collector/pdata/plog"
"go.opentelemetry.io/collector/pdata/ptrace"
)
// maxIssueTitleLen bounds how much of a log body is kept as an issue title,
// so a single noisy log line can't blow up issue payload size.
const maxIssueTitleLen = 200
// extractSpanErrors scans traces for spans with an error status and records
// one issue occurrence per error found.
func extractSpanErrors(tracker *issues.Tracker, td ptrace.Traces, timestamp int64) {
resourceSpans := td.ResourceSpans()
for i := 0; i < resourceSpans.Len(); i++ {
rs := resourceSpans.At(i)
service := resourceServiceName(rs.Resource())
scopeSpans := rs.ScopeSpans()
for j := 0; j < scopeSpans.Len(); j++ {
spans := scopeSpans.At(j).Spans()
for k := 0; k < spans.Len(); k++ {
span := spans.At(k)
if span.Status().Code() != ptrace.StatusCodeError {
continue
}
title := spanErrorTitle(span)
tracker.Record("span_error", service, title, issues.Sample{
Timestamp: timestamp,
TraceID: idString(span.TraceID()),
SpanID: idString(span.SpanID()),
Message: title,
})
}
}
}
}
// extractLogErrors scans logs for error/fatal severity records and records
// one issue occurrence per error found.
func extractLogErrors(tracker *issues.Tracker, ld plog.Logs, timestamp int64) {
resourceLogs := ld.ResourceLogs()
for i := 0; i < resourceLogs.Len(); i++ {
rl := resourceLogs.At(i)
service := resourceServiceName(rl.Resource())
scopeLogs := rl.ScopeLogs()
for j := 0; j < scopeLogs.Len(); j++ {
records := scopeLogs.At(j).LogRecords()
for k := 0; k < records.Len(); k++ {
record := records.At(k)
if !isErrorLog(record) {
continue
}
title := logErrorTitle(record)
tracker.Record("log_error", service, title, issues.Sample{
Timestamp: timestamp,
TraceID: idString(record.TraceID()),
SpanID: idString(record.SpanID()),
Message: title,
})
}
}
}
}
func isErrorLog(record plog.LogRecord) bool {
if record.SeverityNumber() >= plog.SeverityNumberError {
return true
}
switch strings.ToUpper(strings.TrimSpace(record.SeverityText())) {
case "ERROR", "FATAL", "CRITICAL":
return true
default:
return false
}
}
func spanErrorTitle(span ptrace.Span) string {
events := span.Events()
for i := 0; i < events.Len(); i++ {
event := events.At(i)
if event.Name() != "exception" {
continue
}
excType := attrString(event.Attributes(), "exception.type")
excMsg := attrString(event.Attributes(), "exception.message")
switch {
case excType != "" && excMsg != "":
return excType + ": " + excMsg
case excMsg != "":
return excMsg
case excType != "":
return excType
}
}
if msg := span.Status().Message(); msg != "" {
return span.Name() + ": " + msg
}
return span.Name() + " (error)"
}
func logErrorTitle(record plog.LogRecord) string {
excType := attrString(record.Attributes(), "exception.type")
excMsg := attrString(record.Attributes(), "exception.message")
switch {
case excType != "" && excMsg != "":
return truncate(excType + ": " + excMsg)
case excMsg != "":
return truncate(excMsg)
}
body := strings.TrimSpace(record.Body().AsString())
if body == "" {
return "log error"
}
return truncate(body)
}
func truncate(s string) string {
if len(s) <= maxIssueTitleLen {
return s
}
return s[:maxIssueTitleLen] + "…"
}
func resourceServiceName(res pcommon.Resource) string {
if name := attrString(res.Attributes(), "service.name"); name != "" {
return name
}
return "unknown"
}
// attrString safely reads a string attribute, since a Value obtained from a
// missing Map key is not valid to call methods on.
func attrString(attrs pcommon.Map, key string) string {
if v, ok := attrs.Get(key); ok {
return v.AsString()
}
return ""
}
type traceOrSpanID interface {
IsEmpty() bool
String() string
}
func idString(id traceOrSpanID) string {
if id.IsEmpty() {
return ""
}
return id.String()
}