Add Task SDK, Go and Java SDK execution architecture diagrams#69750
Merged
potiuk merged 2 commits intoJul 19, 2026
Conversation
jason810496
approved these changes
Jul 11, 2026
jason810496
left a comment
jason810496
left a comment
Member
There was a problem hiding this comment.
Thanks! It's really a cool one to showcase the current architecture for Lang-SDK, no major error to fix, and we can improve in the follow-up.
potiuk
force-pushed
the
docs-task-execution-architecture-diagrams
branch
2 times, most recently
from
July 11, 2026 16:24
b533189 to
42ee2a6
Compare
Add graphviz-generated diagrams to the architecture overview showing how a task actually runs across the different SDKs, and embed them in core-concepts/overview.rst: - Python Task SDK: supervised (two native OS processes over a msgpack socket) vs native/in-process paths, plus a numbered message-flow sequence. - Go SDK: standalone edge worker pulling from the Edge API, task talks to the Execution API directly. - Java (JVM) SDK: Coordinator layer reusing the Python Supervisor, driving a JVM subprocess over loopback TCP, with an architecture diagram and a numbered execution-workflow sequence.
The message-flow diagrams added earlier in this PR drew every step on a single vertical spine, so the request/response round-trips between the task runtime and the Supervisor were hard to follow. Responding to review feedback, redraw them as true sequence diagrams — one lifeline per participant with the Supervisor in the middle — so each round-trip reads as arrows going back and forth between neighboring lifelines, with straight message arrows and margined labels.
potiuk
force-pushed
the
docs-task-execution-architecture-diagrams
branch
from
July 18, 2026 17:16
42ee2a6 to
10c2991
Compare
potiuk
changed the title
Contributor
Backport successfully created: v3-3-testNote: As of Merging PRs targeted for Airflow 3.X In matter of doubt please ask in #release-management Slack channel.
|
github-actions Bot
pushed a commit
to aws-mwaa/upstream-to-airflow
that referenced
this pull request
Jul 19, 2026
…rams (apache#69750) * Add Task SDK, Go and Java SDK execution architecture diagrams Add graphviz-generated diagrams to the architecture overview showing how a task actually runs across the different SDKs, and embed them in core-concepts/overview.rst: - Python Task SDK: supervised (two native OS processes over a msgpack socket) vs native/in-process paths, plus a numbered message-flow sequence. - Go SDK: standalone edge worker pulling from the Edge API, task talks to the Execution API directly. - Java (JVM) SDK: Coordinator layer reusing the Python Supervisor, driving a JVM subprocess over loopback TCP, with an architecture diagram and a numbered execution-workflow sequence. * Draw Task and Java SDK execution flows as true sequence diagrams The message-flow diagrams added earlier in this PR drew every step on a single vertical spine, so the request/response round-trips between the task runtime and the Supervisor were hard to follow. Responding to review feedback, redraw them as true sequence diagrams — one lifeline per participant with the Supervisor in the middle — so each round-trip reads as arrows going back and forth between neighboring lifelines, with straight message arrows and margined labels. (cherry picked from commit 2de0223) Co-authored-by: Jarek Potiuk <jarek@potiuk.com>
aws-airflow-bot
pushed a commit
to aws-mwaa/upstream-to-airflow
that referenced
this pull request
Jul 19, 2026
…rams (apache#69750) * Add Task SDK, Go and Java SDK execution architecture diagrams Add graphviz-generated diagrams to the architecture overview showing how a task actually runs across the different SDKs, and embed them in core-concepts/overview.rst: - Python Task SDK: supervised (two native OS processes over a msgpack socket) vs native/in-process paths, plus a numbered message-flow sequence. - Go SDK: standalone edge worker pulling from the Edge API, task talks to the Execution API directly. - Java (JVM) SDK: Coordinator layer reusing the Python Supervisor, driving a JVM subprocess over loopback TCP, with an architecture diagram and a numbered execution-workflow sequence. * Draw Task and Java SDK execution flows as true sequence diagrams The message-flow diagrams added earlier in this PR drew every step on a single vertical spine, so the request/response round-trips between the task runtime and the Supervisor were hard to follow. Responding to review feedback, redraw them as true sequence diagrams — one lifeline per participant with the Supervisor in the middle — so each round-trip reads as arrows going back and forth between neighboring lifelines, with straight message arrows and margined labels. (cherry picked from commit 2de0223) Co-authored-by: Jarek Potiuk <jarek@potiuk.com>
potiuk
added a commit
that referenced
this pull request
Jul 19, 2026
…rams (#69750) (#70100) * Add Task SDK, Go and Java SDK execution architecture diagrams Add graphviz-generated diagrams to the architecture overview showing how a task actually runs across the different SDKs, and embed them in core-concepts/overview.rst: - Python Task SDK: supervised (two native OS processes over a msgpack socket) vs native/in-process paths, plus a numbered message-flow sequence. - Go SDK: standalone edge worker pulling from the Edge API, task talks to the Execution API directly. - Java (JVM) SDK: Coordinator layer reusing the Python Supervisor, driving a JVM subprocess over loopback TCP, with an architecture diagram and a numbered execution-workflow sequence. * Draw Task and Java SDK execution flows as true sequence diagrams The message-flow diagrams added earlier in this PR drew every step on a single vertical spine, so the request/response round-trips between the task runtime and the Supervisor were hard to follow. Responding to review feedback, redraw them as true sequence diagrams — one lifeline per participant with the Supervisor in the middle — so each round-trip reads as arrows going back and forth between neighboring lifelines, with straight message arrows and margined labels. (cherry picked from commit 2de0223) Co-authored-by: Jarek Potiuk <jarek@potiuk.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds graphviz-generated diagrams to the architecture overview
(core-concepts/overview.rst) explaining how a task actually runs across
Airflow's SDKs, in a new "Task execution architecture" section:
msgpack socket) vs the native in-process path (dag.test()), plus a numbered
message-flow sequence.
the compiled bundle as a go-plugin gRPC subprocess and calling the Execution
API directly.
-> _JavaActivitySubprocess) reusing the Python Supervisor, driving a JVM
subprocess over loopback TCP, with an architecture diagram and a numbered
execution-workflow sequence.
All diagrams are generated by the existing generate-airflow-diagrams prek hook
(committed as .py + .png + .md5sum).
Was generative AI tooling used to co-author this PR?
Generated-by: Claude Code (Opus 4.8) following the guidelines