- 243 PRs by 61 committers
- 437 messages on the mailing list
- 40 companies
- Upcoming unsolicited talk: "A Practical Introduction to Airflow" at PyBay by MattDavis (Clover health)
No wonder it's hard to keep up!
| Author | prs |
|---|---|
| jlowin | 38 |
| Bolke de Bruin | 28 |
| Chris Riccomini | 21 |
| Dan Davydov | 18 |
| Maxime Beauchemin | 11 |
| Kengo Seki | 11 |
| Siddharth Anand | 11 |
| Stanislav Kudriashev | 8 |
| Sid Anand | 7 |
| Sumit Maheshwari | 7 |
| Arthur Wiedmer | 6 |
| Li Xuanji | 5 |
| Alex Van Boxel | 4 |
| Eric Stern | 4 |
| Hongbo Zeng | 3 |
| Hervé Werner | 3 |
| Norman Mu | 3 |
| Alexey Ustyantsev | 3 |
| Ajay Yadav | 2 |
| Rob Froetscher | 2 |
| Ilya Rakoshes | 2 |
| Tsuyoshi Ozawa | 2 |
| John Bodley | 2 |
| Joy Gao | 2 |
| Junwei Wang | 2 |
| Yap Sok Ann | 2 |
- Scheduler complete rewrite with amazing cross company collaboration
- Scheduler is threaded and will scale to the next order of magnitude
- Scheduler is insulated, DAG code is parsed in subprocesses
- jlowin's Git integration tool makes working with Apache's constraint so much easier, less friction when merging is really important
- @mistercrunch no longer a bottleneck for the project, many engineers from different companies know the core of the project
- More logging & stats collection in critical areas
- Refactor of the dependency engine (@aoen pending merge), along with a "Why isn't my task triggered" web page
- Workload management for Hive
- SparkSql & EMR operators
- Apache compliance (removing Highcharts, license headers)
- Tons of bug fixes, tweaks and polish