#
# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements. See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership. The ASF licenses this file
# to you under the Apache License, Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on an
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied. See the License for the
# specific language governing permissions and limitations
# under the License.
from __future__ import annotations
import os
import pathlib
import shutil
from packaging.version import Version
from airflow.compat.functools import cached_property
from airflow.configuration import conf
from airflow.providers.amazon.aws.hooks.s3 import S3Hook
from airflow.utils.log.file_task_handler import FileTaskHandler
from airflow.utils.log.logging_mixin import LoggingMixin
[docs]def get_default_delete_local_copy():
"""Load delete_local_logs conf if Airflow version > 2.6 and return False if not
TODO: delete this function when min airflow version >= 2.6
"""
from airflow.version import version
if Version(version) < Version("2.6"):
return False
return conf.getboolean("logging", "delete_local_logs")
[docs]class S3TaskHandler(FileTaskHandler, LoggingMixin):
"""
S3TaskHandler is a python log handler that handles and reads
task instance logs. It extends airflow FileTaskHandler and
uploads to and reads from S3 remote storage.
"""
[docs] trigger_should_wrap = True
def __init__(
self, base_log_folder: str, s3_log_folder: str, filename_template: str | None = None, **kwargs
):
super().__init__(base_log_folder, filename_template)
self.remote_base = s3_log_folder
self.log_relative_path = ""
self._hook = None
self.closed = False
self.upload_on_close = True
self.delete_local_copy = (
kwargs["delete_local_copy"] if "delete_local_copy" in kwargs else get_default_delete_local_copy()
)
@cached_property
[docs] def hook(self):
"""Returns S3Hook."""
return S3Hook(
aws_conn_id=conf.get("logging", "REMOTE_LOG_CONN_ID"), transfer_config_args={"use_threads": False}
)
[docs] def set_context(self, ti):
super().set_context(ti)
# Local location and remote location is needed to open and
# upload local log file to S3 remote storage.
full_path = self.handler.baseFilename
self.log_relative_path = pathlib.Path(full_path).relative_to(self.local_base).as_posix()
is_trigger_log_context = getattr(ti, "is_trigger_log_context", False)
self.upload_on_close = is_trigger_log_context or not ti.raw
# Clear the file first so that duplicate data is not uploaded
# when re-using the same path (e.g. with rescheduled sensors)
if self.upload_on_close:
with open(self.handler.baseFilename, "w"):
pass
[docs] def close(self):
"""Close and upload local log file to remote storage S3."""
# When application exit, system shuts down all handlers by
# calling close method. Here we check if logger is already
# closed to prevent uploading the log to remote storage multiple
# times when `logging.shutdown` is called.
if self.closed:
return
super().close()
if not self.upload_on_close:
return
local_loc = os.path.join(self.local_base, self.log_relative_path)
remote_loc = os.path.join(self.remote_base, self.log_relative_path)
if os.path.exists(local_loc):
# read log and remove old logs to get just the latest additions
log = pathlib.Path(local_loc).read_text()
write_to_s3 = self.s3_write(log, remote_loc)
if write_to_s3 and self.delete_local_copy:
shutil.rmtree(os.path.dirname(local_loc))
# Mark closed so we don't double write if close is called twice
self.closed = True
def _read_remote_logs(self, ti, try_number, metadata=None):
# Explicitly getting log relative path is necessary as the given
# task instance might be different than task instance passed in
# in set_context method.
worker_log_rel_path = self._render_filename(ti, try_number)
logs = []
messages = []
bucket, prefix = self.hook.parse_s3_url(s3url=os.path.join(self.remote_base, worker_log_rel_path))
keys = self.hook.list_keys(bucket_name=bucket, prefix=prefix)
if keys:
keys = [f"s3://{bucket}/{key}" for key in keys]
messages.extend(["Found logs in s3:", *[f" * {x}" for x in sorted(keys)]])
for key in sorted(keys):
logs.append(self.s3_read(key, return_error=True))
else:
messages.append(f"No logs found on s3 for ti={ti}")
return messages, logs
def _read(self, ti, try_number, metadata=None):
"""
Read logs of given task instance and try_number from S3 remote storage.
If failed, read the log from task instance host machine.
todo: when min airflow version >= 2.6 then remove this method (``_read``)
:param ti: task instance object
:param try_number: task instance try_number to read logs from
:param metadata: log metadata,
can be used for steaming log reading and auto-tailing.
"""
# from airflow 2.6 we no longer implement the _read method
if hasattr(super(), "_read_remote_logs"):
return super()._read(ti, try_number, metadata)
# if we get here, we're on airflow < 2.6 and we use this backcompat logic
messages, logs = self._read_remote_logs(ti, try_number, metadata)
if logs:
return "".join(f"*** {x}\n" for x in messages) + "\n".join(logs), {"end_of_log": True}
else:
if metadata and metadata.get("log_pos", 0) > 0:
log_prefix = ""
else:
log_prefix = "*** Falling back to local log\n"
local_log, metadata = super()._read(ti, try_number, metadata)
return f"{log_prefix}{local_log}", metadata
[docs] def s3_log_exists(self, remote_log_location: str) -> bool:
"""
Check if remote_log_location exists in remote storage
:param remote_log_location: log's location in remote storage
:return: True if location exists else False
"""
return self.hook.check_for_key(remote_log_location)
[docs] def s3_read(self, remote_log_location: str, return_error: bool = False) -> str:
"""
Returns the log found at the remote_log_location. Returns '' if no
logs are found or there is an error.
:param remote_log_location: the log's location in remote storage
:param return_error: if True, returns a string error message if an
error occurs. Otherwise returns '' when an error occurs.
:return: the log found at the remote_log_location
"""
try:
return self.hook.read_key(remote_log_location)
except Exception as error:
msg = f"Could not read logs from {remote_log_location} with error: {error}"
self.log.exception(msg)
# return error if needed
if return_error:
return msg
return ""
[docs] def s3_write(self, log: str, remote_log_location: str, append: bool = True, max_retry: int = 1) -> bool:
"""
Writes the log to the remote_log_location and return `True` when done. Fails silently
and return `False` if no log was created.
:param log: the log to write to the remote_log_location
:param remote_log_location: the log's location in remote storage
:param append: if False, any existing log file is overwritten. If True,
the new log is appended to any existing logs.
:param max_retry: Maximum number of times to retry on upload failure
:return: whether the log is successfully written to remote location or not.
"""
try:
if append and self.s3_log_exists(remote_log_location):
old_log = self.s3_read(remote_log_location)
log = "\n".join([old_log, log]) if old_log else log
except Exception:
self.log.exception("Could not verify previous log to append")
return False
# Default to a single retry attempt because s3 upload failures are
# rare but occasionally occur. Multiple retry attempts are unlikely
# to help as they usually indicate non-ephemeral errors.
for try_num in range(1 + max_retry):
try:
self.hook.load_string(
log,
key=remote_log_location,
replace=True,
encrypt=conf.getboolean("logging", "ENCRYPT_S3_LOGS"),
)
break
except Exception:
if try_num < max_retry:
self.log.warning("Failed attempt to write logs to %s, will retry", remote_log_location)
else:
self.log.exception("Could not write logs to %s", remote_log_location)
return False
return True
