nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Confluence table to JIRA creation

Ankit Gupta_bhakzs

Hi,

I am trying to solve a use case, where I have requirements written in a table in confluence page, and on transition of the page to verified status, the automation shall trigger creation of JIRA tasks 1 per table requirement row. I tried using custom Rovo agents, but catch here is my table has 150+ rows/entries, and agent is unable to parse this much data.
I tried another way to automate the flow and fetch the table data via HTTP GET request, but that provides data which is not easily parsed.
I also tried to convert the Confluence table to database entries using Rovo agent but no luck
Would appreciate any suggestions in this regard.

Thanks in Advance!

Accepted answers

All comments

Swati Rastogi_7gmb9n

Curious if you can add additional filters and make multiple tables that can be read individually and it can then parse in smaller sub-sections?

Tomislav Tobijas

Hi @Ankit Gupta ,

Does this need to be a one-time thing, or do you need to dynamically create new Jira items from the table?

Those larger tables (or content in general) can be a bit tricky. If you're open to manually doing the conversion, you could simply select the cells/rows and then use native integration and option to create Jira work items from those:

If we circle back, I do believe a database would be a more appropriate type for this. What I've found could be a good way to create a database, and particularly for your case, is to potentially export your current page to PDF. Then, use that file together with create with Rovo feature and ask Rovo to create a database. This was the most stable approach I've seen so far 👀

As for the agent itself, you could maybe try the following:

Instructional Batching: Update your agent's instructions to process data in segments. For example: "Analyze the first 30 rows of the table, create tasks, then wait for my confirmation to proceed to the next 30" .

Reasoning Tier: Ensure the agent is set to the "Think deeper" reasoning tier, which allows for up to 60 iterative loops compared to the standard 10

There are a couple of JAC tickets around this:

Apparently, Atlassian suggests defining the exact table structure in the agent's instructions and providing a small example to improve reliability.

Hope this helps.

Cheers,
Tobi

Ankit Gupta_bhakzs

Thanks @Tomislav Tobijas for the insights and quick response.

My intention is to have the database table/Confluence table created which can be updated/modified. Using automation/AI Agent to parse the table (from database or confluence page) and create the JIRA tasks. I have already created a database but seems like in automation the database table rows are unavailable via Rest API. Because of this the automation is not working. And when using same data from a table on a confluence page, the results are too complex to be parsed.
I had tried giving instructions to the Rovo agent by process in batches but somehow, it is unable to do so, in terms of creating the tasks.

Looking forward to your response!

Ankit Gupta_bhakzs

Hi @Swati Rastogi There could be multiple pages/tables already and intention is to automate this dynamic process of whenever a page transitions to a state, read the confluence page table/database table parse the data per row basis and create JIRA tasks.

Regards

Tomislav Tobijas

@Ankit Gupta yeah... REST API for databases is fairly limited. There's a highly voted feature request to expand API endpoints for further manipulation of entries > see CONFCLOUD-77328: Allow Manipulation of Confluence Databases using the REST API

As for Rovo, you could maybe try using subagents, but I'm not sure if that will work. For example, in those cases where you have large tables, you could say that each subagent checks one particular part of the table (e.g. 1st does the first 50 rows, the second checks from 51 to 100, etc.)
But I'm still skeptical if that will actually work 🤔

Apart from actually splitting those into separate datasets/pages, I don't see any stable workaround.

Ankit Gupta_bhakzs

Yes. I have earlier tried using sub-agents as well. No luck, what happens with sub-agents is once there is need for the agent to confirm the creation of JIRAs for the first batch it does not proceed with the other sub-agent's request.

Mia Tamm _Simpleasyty_

Hi @Ankit Gupta I think the key question here is whether those 150 rows should really continue being treated as page content.

If every row represents an independent requirement that later needs automation, status and a Jira work item, modelling each row as structured data would probably make the workflow much more reliable than asking an agent to repeatedly interpret a large rendered table.

I’d also think about idempotency early: how will the automation know that row 47 already created a Jira item and should not create it again?

Welcome!

It looks like you're new here. Sign in or register to get started.

Quick Links

Find more posts tagged with

confluence

jira

cloud

confluence-cloud

automation

rovo