airflow.providers.google.cloud.log.gcs_task_handler
¶
Module Contents¶
-
class
airflow.providers.google.cloud.log.gcs_task_handler.
GCSTaskHandler
(*, base_log_folder: str, gcs_log_folder: str, filename_template: str, gcp_key_path: Optional[str] = None, gcp_keyfile_dict: Optional[dict] = None, gcp_scopes: Optional[Collection[str]] = _DEFAULT_SCOPESS, project_id: Optional[str] = None)[source]¶ Bases:
airflow.utils.log.file_task_handler.FileTaskHandler
,airflow.utils.log.logging_mixin.LoggingMixin
GCSTaskHandler is a python log handler that handles and reads task instance logs. It extends airflow FileTaskHandler and uploads to and reads from GCS remote storage. Upon log reading failure, it reads from host machine's local disk.
- Parameters
base_log_folder (str) -- Base log folder to place logs.
gcs_log_folder (str) -- Path to a remote location where logs will be saved. It must have the prefix
gs://
. For example:gs://bucket/remote/log/location
filename_template (str) -- template filename string
gcp_key_path (str) -- Path to Google Cloud Service Account file (JSON). Mutually exclusive with gcp_keyfile_dict. If omitted, authorization based on the Application Default Credentials will be used.
gcp_keyfile_dict (dict) -- Dictionary of keyfile parameters. Mutually exclusive with gcp_key_path.
gcp_scopes (str) -- Comma-separated string containing OAuth2 scopes
project_id (str) -- Project ID to read the secrets from. If not passed, the project ID from credentials will be used.
-
_read
(self, ti, try_number, metadata=None)[source]¶ Read logs of given task instance and try_number from GCS. If failed, read the log from task instance host machine.
- Parameters
ti -- task instance object
try_number -- task instance try_number to read logs from
metadata -- log metadata, can be used for steaming log reading and auto-tailing.