Manages a project-level logging sink. For more information see the official documentation, Exporting Logs in the API and API.

Note that you must have the "Logs Configuration Writer" IAM role (roles/logging.configWriter) granted to the credentials used with terraform.

Example Usage

resource "google_logging_project_sink" "my-sink" {
    name = "my-pubsub-instance-sink"

    # Can export to pubsub, cloud storage, or bigtable
    destination = "pubsub.googleapis.com/projects/my-project/topics/instance-activity"

    # Log all WARN or higher severity messages relating to instances
    filter = "resource.type = gce_instance AND severity >= WARN"

    # Use a unique writer (creates a unique service account used for writing)
    unique_writer_identity = true

A more complete example follows: this creates a compute instance, as well as a log sink that logs all activity to a cloud storage bucket. Because we are using unique_writer_identity, we must grant it access to the bucket. Note that this grant requires the "Project IAM Admin" IAM role (roles/resourcemanager.projectIamAdmin) granted to the credentials used with terraform.

# Our logged compute instance
resource "google_compute_instance" "my-logged-instance" {
  name         = "my-instance"
  machine_type = "n1-standard-1"
  zone         = "us-central1-a"

  boot_disk {
    initialize_params {
      image = "debian-cloud/debian-8"

  network_interface {
    network = "default"

    access_config {}

# A bucket to store logs in
resource "google_storage_bucket" "log-bucket" {
    name     = "my-unique-logging-bucket"

# Our sink; this logs all activity related to our "my-logged-instance" instance
resource "google_logging_project_sink" "instance-sink" {
    name = "my-instance-sink"
    destination = "storage.googleapis.com/${google_storage_bucket.log-bucket.name}"
    filter = "resource.type = gce_instance AND resource.labels.instance_id = \"${google_compute_instance.my-logged-instance.instance_id}\""

    unique_writer_identity = true

# Because our sink uses a unique_writer, we must grant that writer access to the bucket.
resource "google_project_iam_binding" "log-writer" {
    role = "roles/storage.objectCreator"

    members = [

Argument Reference

The following arguments are supported:

  • name - (Required) The name of the logging sink.

  • destination - (Required) The destination of the sink (or, in other words, where logs are written to). Can be a Cloud Storage bucket, a PubSub topic, or a BigQuery dataset. Examples: "storage.googleapis.com/[GCS_BUCKET]" "bigquery.googleapis.com/projects/[PROJECT_ID]/datasets/[DATASET]" "pubsub.googleapis.com/projects/[PROJECT_ID]/topics/[TOPIC_ID]" The writer associated with the sink must have access to write to the above resource.

  • filter - (Optional) The filter to apply when exporting logs. Only log entries that match the filter are exported. See Advanced Log Filters for information on how to write a filter.

  • project - (Optional) The ID of the project to create the sink in. If omitted, the project associated with the provider is used.

  • unique_writer_identity - (Optional) Whether or not to create a unique identity associated with this sink. If false (the default), then the writer_identity used is serviceAccount:[email protected]. If true, then a unique service account is created and used for this sink. If you wish to publish logs across projects, you must set unique_writer_identity to true.

Attributes Reference

In addition to the arguments listed above, the following computed attributes are exported:

  • writer_identity - The identity associated with this sink. This identity must be granted write access to the configured destination.


Project-level logging sinks can be imported using their URI, e.g.

$ terraform import google_logging_project_sink.my_sink projects/my-project/sinks/my-sink

© 2018 HashiCorpLicensed under the MPL 2.0 License.