Ability to retry child workflow tasks from parent workflow
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 6.5k
- Forks
- 787
- PR merge metrics
- No merged PRs in 30d
Description
Problem:
We cannot re-run a child workflow task from a parent workflow.
Description
Often times a parent workflow will only have 3 tasks. The child workflow will have up to 20 or more tasks.
When the child workflow fails at task 19 we are only able to restart the child workflow at task 20. This will not continue the Parent workflow at task2.
Why are our workflows structured like this? Often times we need to slightly alter some logic for sensor based workflow calls vs manual based workflow calls. The child workflow allows the code to be reused in both parent workflows for sensor and manual runs.
Why do we want to rerun? Sometimes network errors lead to remote ssh runners failing. Rerunning from the parent task1 would do all the earlier steps and duplicate work.
Stackstorm has the ability to restart a workflow from a specific task using the syntax st2 execution re-run <execution-id> --tasks <task_name>
https://docs.stackstorm.com/orquesta/operations.html#re-running-workflow-execution-from-task-s
Proposed UI solution:
allow ability to re-run child workflow tasks like so
st2 execution re-run <execution-id> --tasks <task_name>.<child_task_name>
Developer Notes
My initial research led me to this section of code and a possible solution
https://github.com/StackStorm/st2/blob/d61b563c418f0f82f339a67053df421c8bbbc6c1/st2common/st2common/services/workflows.py#L505
1.tasks = split task on dot
2. find tasks[0] task in workflow
3. if tasks[1] exists then set this task as running and recurse into this function using tasks[1] and workflow from tasks[0] if it is a workflow
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start with st2common/services/workflows.py around line 505 and review the existing task re-run behavior described in the Orquesta operations documentation. Trace how a dotted parent.child task path could locate the child workflow and recurse into it. Done means a child task can be re-run from the parent execution while preserving continuation and avoiding duplicate earlier work.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- backend, devops
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100