Skip to main content

Observability

Remembr can export OpenTelemetry traces and metrics to any OTLP-compatible collector.

Enable it​

OTEL_ENABLED=true
OTEL_EXPORTER_ENDPOINT=http://otel-collector:4317
OTEL_SERVICE_NAME=remembr
OTEL_TRACES_SAMPLE_RATE=1.0

What Remembr emits​

Traces:

  • FastAPI request spans
  • SQLAlchemy database spans
  • Redis spans
  • HTTPX client spans
  • Celery worker spans
  • Custom spans such as memory.search, embedding.generate, and webhook.deliver

Metrics:

  • remembr_memories_stored_total{org_id}
  • remembr_searches_total{org_id,mode}
  • remembr_embeddings_generated_total{provider,model,status}
  • remembr_webhooks_delivered_total{event,status}
  • remembr_search_duration_ms{mode}
  • remembr_embedding_generation_duration_ms{provider}
  • remembr_embedding_queue_depth
  • remembr_cache_hit_rate
  • remembr_memory_count{org_id}

Collector example​

receivers:
  otlp:
    protocols:
      grpc:
      http:

processors:
  batch:

exporters:
  debug:
  prometheusremotewrite:
    endpoint: http://grafana:9090/api/v1/write

service:
  pipelines:
    traces:
      receivers: [otlp]
      processors: [batch]
      exporters: [debug]
    metrics:
      receivers: [otlp]
      processors: [batch]
      exporters: [debug, prometheusremotewrite]

What to watch​

  • Search latency by mode
  • Worker backlog via remembr_embedding_queue_depth
  • Error rate for webhook delivery
  • Cache hit rate during hot traffic windows

Notes​

  • If OTEL_ENABLED=false, Remembr does not configure exporters.
  • OTLP gRPC usually targets port 4317.
  • Queue depth is derived from Redis-backed task state.