generated: '2026-09-12'
method: derived
source: >-
  discovery/google-cloud-logging-discovery-v2.json (revision 20260818, 94 schemas) and
  grpc/google-cloud-logging-logging-config.proto, cross-checked against the REST reference at
  https://docs.cloud.google.com/logging/docs/reference/v2/rest
description: >-
  The entity graph of Cloud Logging v2. Two halves: the telemetry plane (a LogEntry written
  into a log name and stored in a bucket) and the configuration plane (buckets, views, sinks,
  exclusions, metrics and links, all owned by a resource container).
resource_name_scheme:
  form: "{container}/{containerId}[/locations/{location}]/{collection}/{id}"
  containers: [projects, folders, organizations, billingAccounts]
  examples:
    - projects/PROJECT_ID/logs/LOG_ID
    - projects/PROJECT_ID/locations/LOCATION/buckets/BUCKET_ID
    - projects/PROJECT_ID/locations/LOCATION/buckets/BUCKET_ID/views/VIEW_ID
    - projects/PROJECT_ID/sinks/SINK_ID
  note: >-
    There is no opaque id-prefix scheme (no cus_/inv_ style ids). Identity is the full
    hierarchical resource name, and every management method addresses resources by it.
entities:
  - name: LogEntry
    plane: telemetry
    key_fields: [logName, timestamp, insertId, resource, severity, trace, spanId]
    payload_oneof: [textPayload, jsonPayload, protoPayload]
    note: >-
      The unit of ingestion. logName is a full resource name; resource is a
      MonitoredResource; insertId is the client-supplied de-duplication key.
  - name: MonitoredResource
    plane: telemetry
    key_fields: [type, labels]
  - name: MonitoredResourceDescriptor
    plane: telemetry
    key_fields: [type, labels, displayName]
  - name: LogBucket
    plane: configuration
    key_fields: [name, retentionDays, locked, lifecycleState, cmekSettings, indexConfigs, analyticsEnabled]
  - name: LogView
    plane: configuration
    key_fields: [name, filter]
  - name: LogSink
    plane: configuration
    key_fields: [name, destination, filter, writerIdentity, disabled, includeChildren, exclusions]
  - name: LogExclusion
    plane: configuration
    key_fields: [name, filter, disabled]
  - name: LogMetric
    plane: configuration
    key_fields: [name, filter, metricDescriptor, valueExtractor, labelExtractors, bucketName]
  - name: Link
    plane: configuration
    key_fields: [name, bigqueryDataset]
    note: A linked BigQuery dataset over an analytics-enabled bucket.
  - name: CmekSettings
    plane: configuration
    key_fields: [kmsKeyName, serviceAccountId]
  - name: Settings
    plane: configuration
    key_fields: [kmsKeyName, storageLocation, disableDefaultSink]
  - name: Operation
    plane: control
    key_fields: [name, metadata, done, error, response]
relationships:
  - from: LogEntry
    to: MonitoredResource
    kind: has_one
    via: resource
  - from: LogEntry
    to: LogBucket
    kind: belongs_to
    via: logName routed by the Log Router into a bucket
    note: Indirect — the Log Router decides the destination bucket; LogEntry carries no bucket field.
  - from: LogEntry
    to: Trace span
    kind: has_one
    via: trace + spanId
  - from: LogBucket
    to: LogView
    kind: has_many
    via: parent resource name
  - from: LogBucket
    to: CmekSettings
    kind: has_one
    via: cmekSettings
  - from: LogBucket
    to: Link
    kind: has_many
    via: parent resource name
  - from: LogSink
    to: LogBucket
    kind: has_one
    via: destination (logging.googleapis.com/projects/.../buckets/...)
    note: >-
      destination is polymorphic — it may instead name a Cloud Storage bucket, a BigQuery
      dataset, a Pub/Sub topic or another project's log bucket.
  - from: LogSink
    to: LogExclusion
    kind: has_many
    via: exclusions
  - from: LogSink
    to: writerIdentity
    kind: has_one
    via: writerIdentity
    note: >-
      A service account Google mints for the sink; it must be granted write access on the
      destination before routing works. This is the single most common sink misconfiguration.
  - from: LogMetric
    to: LogBucket
    kind: belongs_to
    via: bucketName
  - from: LogMetric
    to: MetricDescriptor
    kind: has_one
    via: metricDescriptor
  - from: Container (project/folder/organization/billingAccount)
    to: [LogBucket, LogSink, LogExclusion, LogMetric, Settings]
    kind: has_many
    via: resource-name parent
  - from: copyLogEntries
    to: Operation
    kind: has_one
    via: long-running operation name