Skip to content

Reconnect Airbyte sync jobs after task retries - #72414

Closed
1fanwang wants to merge 3 commits into
apache:mainfrom
1fanwang:stewang-linkedin-airbyte-resumability
Closed

Reconnect Airbyte sync jobs after task retries#72414
1fanwang wants to merge 3 commits into
apache:mainfrom
1fanwang:stewang-linkedin-airbyte-resumability

Conversation

@1fanwang

@1fanwang 1fanwang commented Sep 2, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

A synchronous Airbyte task can crash after Airbyte accepts a sync, then start a duplicate sync when Airflow retries it. With this change, retries resume active Airbyte jobs and recover completed jobs. A new job is created only when no durable state exists or the previous job failed or was cancelled.

The operator uses AIP-103 ResumableJobMixin only for the synchronous, non-deferrable wait path. Recovery restores job_id before polling or returning so task results, the Airbyte job link, and on_kill() keep pointing at the recovered job. Asynchronous and deferrable execution are unchanged. Airflow versions before 3.3 retain the existing submit behavior.

Testing

Scenario Result
Provider behavior matrix Sync recovery, async/deferrable isolation, on_kill(), and pre-3.3 compatibility covered
Active job stored before retry Reconnected to job 77 without submitting
Current upstream/main with the same stored job Submitted duplicate job 101
Slow deferrable submission Trigger deadlines start after submission
Airflow 2.11 compatibility probe Existing submit behavior preserved
Commands and raw logs 
$ AIRFLOW_HOME=/tmp/airbyte-tests .venv/bin/uv run --project providers/airbyte pytest providers/airbyte/tests/unit/airbyte -q
collected 65 items
...
======================== 65 passed, 1 warning in 28.24s ========================

The recovery probe calls the real operator execute() path with a task-state store containing airbyte_job_id=77. A local stand-in AirbyteHook reports that job as running and records submissions, status checks, and waits.

$ REPRO_VARIANT=upstream/main \
  AIRBYTE_PROVIDER_SOURCE_ROOT=$UPSTREAM_MAIN/providers/airbyte/src \
  .venv/bin/uv run --project providers/airbyte python airbyte_resumable_repro.py
{'variant': 'upstream/main', 'stored_before': 77, 'task_state_get_calls': [], 'task_state_set_calls': [], 'submit_count': 1, 'status_calls': [], 'wait_calls': [101], 'result': 101, 'operator_job_id': 101}
Traceback (most recent call last):
  File "airbyte_resumable_repro.py", line 90, in <module>
    raise RuntimeError("operator did not reconnect to the stored active job")
RuntimeError: operator did not reconnect to the stored active job

$ REPRO_VARIANT="PR branch" \
  AIRBYTE_PROVIDER_SOURCE_ROOT=$PWD/providers/airbyte/src \
  .venv/bin/uv run --project providers/airbyte python airbyte_resumable_repro.py
2026-09-02T05:01:37.864774Z [info     ] Reconnecting to existing job   [airflow.task.operators.airflow.providers.airbyte.operators.airbyte.AirbyteTriggerSyncOperator] external_id=77 external_id_key=airbyte_job_id loc=resumablejobmixin.py:161 status=running
{'variant': 'PR branch', 'stored_before': 77, 'task_state_get_calls': ['airbyte_job_id'], 'task_state_set_calls': [], 'submit_count': 0, 'status_calls': [77], 'wait_calls': [77], 'result': 77, 'operator_job_id': 77}
Stand-in hook reproducer

Save this as airbyte_resumable_repro.py before running the commands above.

from __future__ import annotations

import os
import sys
from importlib import import_module, invalidate_caches
from pathlib import Path
from types import SimpleNamespace
from typing import Any

from airbyte_api.models import JobStatusEnum

import airflow.providers

if source_root := os.environ.get("AIRBYTE_PROVIDER_SOURCE_ROOT"):
    airflow.providers.__path__.insert(0, str(Path(source_root) / "airflow" / "providers"))
    for module_name in tuple(sys.modules):
        if module_name == "airflow.providers.airbyte" or module_name.startswith(
            "airflow.providers.airbyte."
        ):
            del sys.modules[module_name]
    invalidate_caches()

airbyte_module = import_module("airflow.providers.airbyte.operators.airbyte")


class TaskStateStore:
    def __init__(self) -> None:
        self.values: dict[str, int] = {"airbyte_job_id": 77}
        self.get_calls: list[str] = []
        self.set_calls: list[tuple[str, int]] = []

    def get(self, key: str) -> int | None:
        self.get_calls.append(key)
        return self.values.get(key)

    def set(self, key: str, value: int) -> None:
        self.set_calls.append((key, value))
        self.values[key] = value


class StandInAirbyteHook:
    def __init__(self) -> None:
        self.submit_count = 0
        self.status_calls: list[int] = []
        self.wait_calls: list[int] = []

    def submit_sync_connection(self, *, connection_id: str) -> Any:
        self.submit_count += 1
        return SimpleNamespace(job_id=101, status=JobStatusEnum.RUNNING)

    def get_job_status(self, *, job_id: int) -> str:
        self.status_calls.append(job_id)
        return JobStatusEnum.RUNNING

    def wait_for_job(self, *, job_id: int, wait_seconds: float, timeout: float) -> None:
        self.wait_calls.append(job_id)


task_store = TaskStateStore()
hook = StandInAirbyteHook()
airbyte_module.AirbyteHook = lambda **kwargs: hook
operator = airbyte_module.AirbyteTriggerSyncOperator(
    task_id="airbyte_live_recovery",
    connection_id="connection-uuid",
    deferrable=False,
    asynchronous=False,
    wait_seconds=0,
)
result = operator.execute(context={"task_state_store": task_store})
print(
    {
        "variant": os.environ["REPRO_VARIANT"],
        "stored_before": 77,
        "task_state_get_calls": task_store.get_calls,
        "task_state_set_calls": task_store.set_calls,
        "submit_count": hook.submit_count,
        "status_calls": hook.status_calls,
        "wait_calls": hook.wait_calls,
        "result": result,
        "operator_job_id": operator.job_id,
    }
)

if (
    hook.submit_count != 0
    or hook.status_calls != [77]
    or hook.wait_calls != [77]
    or result != 77
    or operator.job_id != 77
):
    raise RuntimeError("operator did not reconnect to the stored active job")
Was generative AI tooling used to co-author this PR?
  • Yes (GitHub Copilot CLI, GPT-5.6 Sol)

Generated-by: GitHub Copilot CLI (GPT-5.6 Sol) following the guidelines

  • Read the Pull Request Guidelines for more information. Note: commit author/co-author name and email in commits become permanently public when merged.
  • For fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
  • When adding dependency, check compliance with the ASF 3rd Party License Policy.
  • For significant user-facing changes create newsfragment: {pr_number}.significant.rst, in airflow-core/newsfragments. You can add this file in a follow-up commit after the PR is created so you know the PR number.

@1fanwang
Reconnect Airbyte sync jobs after task retries
acf0c06 
Worker crashes during synchronous Airbyte polling can otherwise submit a duplicate sync when the task retries.

Generated-by: GitHub Copilot CLI (GPT-5.6 Sol)
Signed-off-by: 1fanwang <[email protected]>
@1fanwang
Exclude Airbyte submission time from deferrable timeout
1aab29f 
Slow Airbyte submission should not reduce how long the trigger can wait for the submitted job.

Generated-by: GitHub Copilot CLI (GPT-5.6 Sol)
Signed-off-by: 1fanwang <[email protected]>
@1fanwang
Fix Airbyte resumability type checking
780d2dd 
The pre-3.3 compatibility fallback could not be type-checked because the mixin methods only exist on concrete operators.

Generated-by: GitHub Copilot CLI (GPT-5.6 Sol)
Signed-off-by: 1fanwang <[email protected]>
This was referenced Sep 25, 2026
Closed
Closed
Closed
Closed
Closed
Closed
Closed
Closed
Closed
Closed
Closed
Closed
Closed
Closed
Closed
@potiuk potiuk added the closed because of open PR limit Closed as a one-time step of introducing the open pull request limit label Sep 25, 2026
@potiuk

potiuk commented Sep 25, 2026

Copy link
Copy Markdown
Member

Hello @1fanwang - thank you for your contributions to Apache Airflow!

The Airflow community has introduced a limit of 5 open pull requests at a time for contributors without write access to the repository. You currently have 34 open pull requests, so - as a one-time step of introducing the limit - we closed the ones where maintainers have not engaged yet:

These pull requests stay open because maintainers are already engaged in them - they count towards your limit:

This is not a judgement of you or of your changes. We never told contributors before that opening many pull requests at once was a problem, so there is nothing to feel bad about - and nothing is lost: your branches, commits and the review history stay where they are.

What we ask you to do is to make your first prioritization decision: choose which of the pull requests above matter most to you, and reopen them (up to 5 open at a time, including the ones still open) with the "Reopen pull request" button or gh pr reopen <PR_NUMBER> --repo apache/airflow. Reopen the ones you are ready to follow through - keep them rebased, respond to review comments and fix failing checks.

While your pull requests are waiting for review, the most valuable thing you can do is help in other ways - reviewing other contributors' pull requests, helping with issues, and taking part in the discussions on the devlist and Slack.

Why we introduced the limit, what it means for you and how to reopen or restore a pull request is explained in https://github.com/apache/airflow/blob/main/contributing-docs/32_open_pull_request_limit.rst.

Drafted-by: Claude Code (Opus 5); reviewed by @potiuk before posting

@potiuk potiuk closed this Sep 25, 2026
This was referenced Sep 25, 2026
Closed
Closed
Closed
Closed
Closed
Closed
This was referenced Sep 25, 2026
Closed
Closed
Closed
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

area:providers closed because of open PR limit Closed as a one-time step of introducing the open pull request limit kind:documentation provider:airbyte

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@1fanwang @potiuk