Repository navigation
Conversation
A checkpoint response without a token means the current invocation must stop checkpointing. The background thread previously treated every such response as execution completion, which could report SUCCEEDED or FAILED while operation work was still in flight. It now reports PENDING. - Add ExecutionSuspendedByService and a token-revoked latch distinct from execution completion. - Exempt a batch carrying the execution's terminal update because that execution is already complete. - Reject newly queued operations and refreshes after token withdrawal, and wake accepted and queued waiters through the suspension path. - Return a bare PENDING response from every otherwise-terminal exit, including oversized-result paths and the CheckpointError exit, which now consults the latch for symmetry with the other exits. - Treat a refresh-only batch exactly like every other non-terminal batch when its response omits the token. - Document ExecutionSuspendedByService in the Raises section of create_checkpoint and schedule_refresh. - Drop ExecutionSuspendedByService's redundant constructor override, identical to SuspendExecution's. - Describe only the SDK-observed condition and resulting behavior in comments and docstrings. - Cover queue guards, deferred refreshes, completion versus withdrawal, oversized results, the CheckpointError exit, and every PENDING exit. Fixes #722
- Executor: skip PENDING-validator rejection when execution.paused - Add 9 new tests (6 unit + 3 e2e) covering pause/resume semantics - Fix 61 pre-existing executor_test.py mocks missing .paused=False (unconfigured Mock attrs are truthy, broke new paused-aware paths) - Update 3 mock-sequencing tests for new _invoke_execution store.load call
Use the existing SuspendExecution control flow for revoked checkpoint tokens and simplify the surrounding comments and tests. Clean up the local test runner pause and resume path, including deferred invocation naming and focused pause/resume coverage.
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
hln33
force-pushed
the
722-feature-exit-gracefully-with-pending-when-a-checkpoint-response-has-no-checkpointtoken
branch
from
October 6, 2026 18:54
8859615 to
aa2d825
Compare
…n-a-checkpoint-response-has-no-checkpointtoken
yaythomas
requested changes
Oct 7, 2026
yaythomas
left a comment
yaythomas
left a comment
Contributor
There was a problem hiding this comment.
Thank you @hln33, this is a careful port of the JS fix, and the state_test.py coverage of the real checkpoint loop is excellent!
Thanks for the follow-up commit for the Codex findings. _invoke_execution no longer reads and saves outside the worker lane, a retried checkpoint no longer receives the withheld token, and the mid-step pause test now fails without the fix (I checked against main's core SDK).
Please could you rename the PR to the conventional commit format, e.g. fix(sdk): answer PENDING on revoked token (your first commit's subject)? The PR title becomes the squash commit message: https://github.com/aws/aws-durable-execution-sdk-python/blob/main/CONTRIBUTING.md#pull-request-title-and-commit-message-format
Maybe this behaviour also warrants a conformance requirement? If so, please could you note it in the description or open an issue in aws-durable-execution-conformance-tests?
This comment has been minimized.
This comment has been minimized.
Add a finite timeout to pause_execution and preserve the pause request when waiting times out. Raise TimeoutError on expiry, matching wait_for_result and wait_for_callback. Forward the timeout through the test runner. Cover timeout validation, worker errors, and pause/resume behavior with unit and end-to-end tests.
- Reuse queue draining across the three checkpoint exit paths. - Preserve overflow-first order and caller-specific errors. - Include the execution ARN in token-revocation warnings. - Count submitted updates without including refresh-only entries. - Test queue draining, waiter errors, and warning accuracy.
Document guard precedence and checkpoint-loop stop conditions. Distinguish terminal completion from token revocation. Explain why a quota-limit failure can return PENDING without changing the execution status from FAILED. Add two unit tests for guards that run before token revocation.
This was referenced Oct 13, 2026
Contributor
Author
|
@yaythomas Thanks for identifying these follow-ups. I opened separate issues for:
The conformance issue calls out the need to agree on a deterministic test mechanism before adding requirements or SDK handlers. |
hln33
changed the title
…ken-revoked # Conflicts: # packages/aws-durable-execution-sdk-python/src/aws_durable_execution_sdk_python/execution.py
This comment has been minimized.
This comment has been minimized.
…n-a-checkpoint-response-has-no-checkpointtoken
Contributor
Codex AI reviewFour confirmed findings:
Static review completed; tests were not executed under the read-only constraints. Reviewed commit |
This branch was successfully deployed
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.
Issue #, if available:
Fixes #722
Description of changes:
This PR updates the SDK to return
PENDINGwhen a checkpoint response omits the next checkpoint token for non-terminal work, instead of treating that response as execution completion.Changes include:
SuspendExecutionpath.PENDINGwhen the token has been revoked.By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.