Processing for the same JIRA ticket via the Rovo agent sometimes succeeds and sometimes fails.

Question

I have configured automation settings for a space to have Rovo analyze ticket content and append the results as a comment.

Although the ticket content is the same, Rovo's analysis process sometimes succeeds and sometimes fails. When a failure occurs, the following comment appears in the audit log:

■ Comment upon failure: Suspended rule has surpassed its time limit.

Could you please check the following points?
・ The cause of the failure and how to address it.
・ Why the process succeeds in some cases but fails in others.
・ I understand this comment to mean a "processing timeout"; is this correct?
(Assuming the above understanding is correct:)
・ What is the timeout limit?
・ Can the timeout issue be prevented by limiting the scope of Rovo's analysis? If so, what methods would you suggest for limiting the scope? (e.g., prompt phrasing, specifying information sources, etc.)

Thank you for your assistance.

yuki.shikano · Accepted Answer

By the way, is it possible to share this ticket with other members?

Tomislav Tobijas · Accepted Answer

Hi @yuki_shikano ,

Just a note that this is a public forum and you've posted a question (which can be accessed by anyone) and not an official support ticket. If you'd like to submit a ticket to the support team, you can use the following channel: Atlassian Support 🎟️
*note that you need to be an app admin and have an app on a paid plan to create a ticket there

Now, to get back to your question...

"Suspended rule has surpassed its time limit" indicates that your automation rule has hit a processing timeout. This occurs when a rule takes too long to complete its execution, causing the system to terminate it to maintain platform performance.

I'm guessing you're using Rovo action in an automation flow? 👀

To sum it up, it might be that the agent action is performing complex analysis that exceeds the allowed execution window. It can also very if the LLM is at capacity or the content requires more "thinking" time -> meaning, rule may tip over the time limit.
Note that there're also some known bugs and suggestions, which might be related in a way to what you're seeing:

* ROVO-852: Rovo Agents invoked from Automation Rules can sometimes hang with SOME ERRORS state 
* ROVO-592: Add retry mechanism when Rovo agents failed to invoke in automation rule due to rate limiting

🤔 What you could do/try...

* Refine the prompt - like, don't ask it to 'analyze everything' but use specific smart values to focus only on relevant fields
* If the analysis is complex, you could try splitting it into multiple smaller, sequential prompts or flows
* Avoid deep research (unless it's essential)

Cheers,
Tobi

Arkadiusz Wroblewski · Accepted Answer

Hello @yuki_shikano

Atlassian does not currently publish the exact timeout duration for an individual Rovo invocation, and it cannot be configured by an administrator.

Reducing the prompt scope may lower the risk, although it cannot completely prevent a transient backend or rate-limiting failure. I would pass only the required fields, for example:

{{issue.summary}} and telling Rovo explicitly not to search extra sources or invoke extra actions. Adding a Log action right before the Rovo step will also help you compare inputs on failed vs. successful runs.

Best,

Arek🤠