How to get auto-scheduler to respect one person per story constraint

[Evgeny Shapiro]

I have a plan set up in Advanced Roadmaps. The plan contains a number of long-running stories (2–6 weeks). I’m using a plan-only Kanban team with two people and a capacity of 80 hours.

The auto-scheduler assigns both people to the same story, shortening its duration and creating an incorrect plan. In addition, all stories are scheduled sequentially with no concurrency. As a result, a story estimated to take 2 weeks is scheduled as taking only 1 week in the plan.

This behavior contradicts the documentation, which states that the scheduler should assign one person per story.

Are there any settings that control this behavior, or is this a known bug (https://jira.atlassian.com/browse/JRACLOUD-97753)?

So far, I’ve checked the following, but none have had any effect:

• Plan-only team vs. a real team
• The story has 2 weeks of remaining time (original estimate: 2 weeks, no time logged)
• The issue is a standard built-in Story at the story level

Any advice would be greatly appreciated.

[Peter_DevSamurai]

Hi @Evgeny Shapiro , 

You’re right, the Advanced Roadmaps docs say the auto-scheduler should effectively treat stories as worked on by one person at a time, so a 2‑week story shouldn’t become 1 week just because there are 2 people on the team.

For reference, you can view the below documents: 

• How capacity is assigned in plans

• Capacity in Kanban teams

So you can try:

1. Pin work to a single assignee first
   Before running Auto-schedule, explicitly assign each story to one person on the team. This usually prevents the scheduler from “pooling” capacity across multiple people on the same story.

2. Check that you’re using time-based estimates on story-level issues
   For Kanban teams, Plans only uses time estimates on story-level issues. Make sure:

   • The issues are standard stories.

   • They have remaining/original estimate in time units.

If, after that, Auto-schedule is still:

• putting two people on the same story, and

• shrinking a 2‑week story to 1 week,

then what you’re seeing matches this known bug you linked: JRACLOUD-97753

But since, you’re on Enterprise, I’d also raise an Atlassian Support ticket and reference that bug, including:

• a link to the plan,

• one or two example stories (with estimates),

• and a screenshot of the team capacity settings.

That way Support can confirm whether you’re hitting the same defect or a configuration edge case.

I hope everything works out for you!

Best regards, 

Peter

[Arkadiusz Wroblewski]

Hello @Evgeny Shapiro 

I think your reading is correct.

I do not see a documented setting to make auto-scheduling truly respect the individual Assignee. Plans mainly works from team-level capacity. At the same time, Atlassian’s current docs also say that for story-level work, the auto-scheduler assumes one person handles one story at a time. So if a 2-week story is being compressed to 1 week just because the team has 2 people, that does not look right. 

That behavior also matches JRACLOUD-97753 pretty closely. That bug is about the autoscheduler assigning too few days to a story and behaving as if multiple users can work on the same story simultaneously. 

My take would be:

- I do not think there is a native setting you missed here
- this looks more like the same defect / limitation
- if you need strict individual-person scheduling today, the practical workaround is still one team per person or manual scheduling

Since you are on Enterprise, please raise that to Atlassian Support and reference JRACLOUD-97753.