# otel_collector.yaml

receivers:
  # ────────────────────────────────────────────────────────────────
  # 1) Host log collection
  # ────────────────────────────────────────────────────────────────
  filelog:
    # Optional: give the receiver a human‑readable name
    name: host_logs
    # Paths to monitor – adjust glob patterns as needed
    include:
      - /var/log/*.log
      - /var/log/syslog
      - /var/log/messages
    # Optional: skip log files that match these patterns
    exclude: []
    # Optional: start reading from the beginning of the file
    start_at: beginning
    # Optional: decode JSON logs if they are JSON‑formatted
    # json:
    #   keys: [timestamp, level, msg]
    #   timestamp_key: timestamp
    #   timestamp_format: RFC3339
    # Optional: throttle log ingestion
    # throttle:
    #   max_per_second: 1000
    # Optional: keep track of processed offsets
    # cache:
    #   file: /var/log/filelog_cache.json

  # ────────────────────────────────────────────────────────────────
  # 2) Docker container log collection
  # ────────────────────────────────────────────────────────────────
  docker:
    # Connect to the local Docker daemon
    host: unix:///var/run/docker.sock
    # By default, the Docker receiver pulls container metrics, logs, and health‑checks.
    # If you only need logs, set the following to true (enabled by default):
    logs: true
    # Optional: enable or disable container log collection for specific containers
    # include: [ "my-app", "api-service" ]
    # exclude: [ "debug-container" ]
    # Optional: set the maximum number of log files to keep open
    # max_open_files: 100

processors:
  # ────────────────────────────────────────────────────────────────
  # 1) Batch processor (recommended for all pipelines)
  # ────────────────────────────────────────────────────────────────
  batch:
    timeout: 5s
    send_batch_size: 5000

  # ────────────────────────────────────────────────────────────────
  # 2) Optional resource processor – add host name to every log record
  # ────────────────────────────────────────────────────────────────
  resource:
    attributes:
      - key: host.name
        from_env: HOSTNAME
        action: upsert
      - key: service.namespace
        value: host-logs
        action: upsert

exporters:
  # ────────────────────────────────────────────────────────────────
  # 1) OTLP exporter – sends logs to the central collector
  # ────────────────────────────────────────────────────────────────
  otlp:
    endpoint: "centralized-collector:4317"
    # If you want insecure connections (e.g., local dev), set insecure true
    tls:
      insecure: true
    # Optional: use HTTP instead of gRPC
    # http:
    #   path: "/v1/logs"

service:
  pipelines:
    logs:
      receivers: [filelog, docker]
      processors: [batch, resource]
      exporters: [otlp]