docs: align experiment step states with actual platform behavior - #198
Merged
Conversation
The step state table claimed RUNNING requires all targets to be executing, while a step is running as soon as its first target execution starts. It also left the canceled-vs-skipped boundary and the "no matching target" failure cause undocumented, so readers could not tell why a step that never ran was reported one way or the other. Spell out that a step is only CANCELED when at least one target execution was running, and that a step whose action was dispatched but never started is SKIPPED together with all its target executions. (ticket 15296)
ManuelGerding
force-pushed
the
docs/15296-step-state-semantics
branch
from
August 4, 2026 09:49
44aff8a to
07f1fab
Compare
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The step state table on Run an Experiment → States described behavior the platform does not have, found while investigating ticket 15296.
What was wrong
RUNNINGclaimed the action is "currently executed on all targets". The platform marks a stepRUNNINGas soon as its first target execution starts (isAnyTargetRunning()). Requiring all targets would be wrong in practice — targets start at different times, and one stalled target would keep a step out ofRUNNINGindefinitely.COMPLETEDgenuinely is all-targets; that asymmetry is intentional and now stated.CANCELED/SKIPPEDleft the boundary implicit, so a reader could not tell why a step that never ran was reported one way or the other.FAILEDlisted only the failed-check cause, omitting that a step also fails when it has no matching targets anymore.Changes
RUNNING: "on at least one target", with the reason.CANCELED: "was running before" spelled out as at least one target execution having started.FAILED: adds the no-matching-targets cause.CANCELEDif at least one target execution wasRUNNING; if none started, the step and all its target executions areSKIPPED, even if the action had already been sent to the agents.The target-execution table needed no wording changes — it already matched the implementation.
The
CANCELED/SKIPPEDwording documented the intended behavior correctly; the platform is being fixed to match it in steadybit/platform#1818. TheRUNNINGandFAILEDrows are pure doc corrections — no behavior change there.https://steadybitgmbh.kanbanize.com/ctrl_board/2/cards/15296