otel.yml

# Define the protocols to receive data for.
# See https://opentelemetry.io/docs/collector/configuration/#receivers
receivers:
  # Configure receiving OTLP data via gRPC on port 4317.
  otlp:
    protocols:
      grpc:
        endpoint: 0.0.0.0:4317

  # Defines a Prometheus configuration set.
  prometheus:
    # Define a set of configurations for scraping by the OpenTelemetry Collector.
    config:
      # The `scrape_configs` section pertains to the Prometheus `scrape_configs`
      # configuration block.
      # See https://prometheus.io/docs/prometheus/latest/configuration/configuration/#scrape_config
      scrape_configs:
        # Scrape Mimir metrics.
        - job_name: 'mimir'
          static_configs:
            - targets: ['mimir:9009']
              labels:
                service: 'mimir'
                group: 'infrastructure'

        # Scrape Loki metrics.
        - job_name: 'loki'
          static_configs:
            - targets: ['loki:3100']
              labels:
                service: 'loki'
                group: 'infrastructure'

        # Scrape Tempo metrics.
        - job_name: 'tempo'
          static_configs:
            - targets: ['tempo:3200']
              labels:
                service: 'tempo'
                group: 'infrastructure'

        # Scrape Grafana metrics.
        - job_name: 'grafana'
          static_configs:
            - targets: ['grafana:3000']
              labels:
                service: 'grafana'
                group: 'infrastructure'

        # Scrape from the Mythical Server and Requester service.
        - job_name: 'mythical'
          scrape_interval: 2s
          static_configs:
            - targets: ['mythical-server:4000']
              labels:
                service: 'mythical-server'
                group: 'mythical'
            - targets: ['mythical-requester:4001']
              labels:
                service: 'mythical-requester'
                group: 'mythical'

        # Scrape from the Beyla Mythical services.
        - job_name: 'beyla-infra'
          scrape_interval: 15s
          static_configs:
            - targets: ['beyla-requester:9090']
              labels:
                service: 'beyla-requester'
                group: 'beyla'
            - targets: ['beyla-server:9090']
              labels:
                service: 'beyla-server'
                group: 'beyla'
            - targets: ['beyla-recorder:9090']
              labels:
                service: 'beyla-recorder'
                group: 'beyla'


# Define processors to process received data.
# See https://opentelemetry.io/docs/collector/configuration/#processors
processors:
  # Use the in-built `batch` processor to batch up data before writing it for export.
  # Use the default values for it.
  batch:


  # The tail sampler processor will only keep traces where spans match the defined policies.
  tail_sampling:
    decision_wait: 30s    # The time to wait for a decision to be made.
    # The following policies follow a logical OR pattern, meaning that if any of the policies match,
    # the trace will be kept. For logical AND, you can use the `and` policy. Every span of a trace is
    # examined by each policy in turn. A match will cause a short-circuit.
    policies: [
      # This policy defines that traces that include spans that contain errors should be kept.
      {
        name: sample-erroring-traces,           # Name of the policy.
        type: status_code,                      # The type must match the type of policy to be used.
        status_code: { status_codes: [ERROR] }  # Only sample traces which have a span containing an error.
      },
      # This policy defines that traces that are over 200ms should be sampled.
      {
        name: sample-long-traces,               # Name of the policy.
        type: latency,                          # The type must match the type of policy to be used.
        latency: { threshold_ms: 200 },         # Only sample traces which are longer than 200ms in duration.
      },
    ]

# Define processors to process received data.
# See https://opentelemetry.io/docs/collector/configuration/#connectors
connectors:
  # The spanmetrics connector is used to output span metrics based on received trace spans.
  spanmetrics:
    namespace: traces.spanmetrics   # Prefix all metrics with `traces.spanmetrics` (this becomes `traces_spanmetrics`).
    # Determine the type of histogram to use for span metrics.
    histogram:
      explicit:                     # Explicit histograms have pre-defined bucket sizes (use default here).
    # Defines additional label dimensions of the metrics from trace span attributes present.
    dimensions:
      - name: http.method
      - name: http.target
      - name: http.status_code
      - name: service.version
    # Ensure exemplars are enabled and sent to the metrics store.
    exemplars:
      enabled: true

  # The servicegraph connector is used to output service node metrics based on received trace spans.
  servicegraph:
      # Defines which exporter the processor will write metrics to.
      metrics_exporter: prometheusremotewrite
      # Defines additional label dimensions of the metrics from trace span attributes present.
      store:                      # Configuration for the in-memory store.
        ttl: 2s                   # Time to wait for an edge to be completed.
        max_items: 200            # Number of edges that will be stored in the storeMap.
      cache_loop: 2m              # The timeout used to clean the cache periodically.
      store_expiration_loop: 10s  # The timeout used to expire old entries from the store periodically.
      # Virtual node peer attributes allow server nodes to be generated where instrumentation isn't present (eg. where
      # service client calls remotely to a service that does not include instrumentation).
      # Service nodes/edges will be generated for any attribute defined.
      virtual_node_peer_attributes:
        - db.name


# Define exporters to data stores.
# See https://opentelemetry.io/docs/collector/configuration/#exporters
# Also see https://github.com/open-telemetry/opentelemetry-collector/tree/main/processor#recommended-processors
exporters:
  # Exporter for sending trace data to Tempo.
  otlp/grafana:
    # Send to the locally running Tempo service.
    endpoint: tempo:4317
    # TLS is not enabled for the instance.
    tls:
      insecure: true

  # Exporter for sending Prometheus data to Mimir.
  prometheusremotewrite:
    # Send to the locally running Mimir service.
    endpoint: http://mimir:9009/api/v1/push
    # TLS is not enabled for the instance.
    tls:
      insecure: true

# Define the full service graph for the OpenTelemetry collector.
service:
  # A pipeline can exist for each of the signals received.
  pipelines:
    # Define the trace pipeline.
    traces:
      # Receive from the `otlp` receiver.
      receivers: [otlp]
      # Use the `batch` processor to process received trace spans.
      processors: [batch]
      # Comment out other `processor` definitions and uncomment the line below to use tail sampling.
      #processors: [tail_sampling, batch]
      # Comment out other `processor` definitions and uncomment the line below to generate service graph metrics
      # from within the OpenTelemetry Collector.
      #processors: [servicegraph, batch]
      # Export to the `otlp/grafana` exporter.
      exporters: [otlp/grafana]
      # Comment out other `exporters` definitions and uncomment the line below to generate span metrics
      # from within the OpenTelemetry Collector as well as exporting traces to Tempo.
      #exporters: [otlp/grafana, spanmetrics]

    # Define the metrics pipeline.
    metrics:
      # Receive metrics from the `prometheus` receiver.
      receivers: [prometheus]
      # Comment out other `receivers` definitions and uncomment the line below to import spanmetrics as well
      # as prometheus metrics.
      #receivers: [prometheus, spanmetrics]
      # Use the `batch` processor to process received metrics.
      processors: [batch]
      # Export to the `prometheusremtotewrite` exporter.
      exporters: [prometheusremotewrite]