Will Analytics have access to Rovo ussage data?

Jeremy Wood

I have been tasked with creating custom rovo ussage reports for my company. Currently i can export rovo ussage events to a CSV that i can them build custom reports from. But it would be ideal if Atlassian Analytics had this connection for me already.

Are there any future plans for this capability? Or how would i go about requesting it from Atlassian?

hamza.chundrigar

Hi @Jeremy Wood

As far as I know, not yet, unfortunately. Rovo usage and credit consumption data isn't currently available through the Atlassian Data Lake, and I haven't seen a confirmed timeline from Atlassian for adding it.

Atlassian currently provides Rovo reporting through Atlassian Administration, but not as queryable tables in the Data Lake.

I believe you have two main options

• Go to AI and Rovo → Trends to see adoption, active users, interactions, and agent activity. Documentation
• Go to Usage → Meters → Rovo credits (or Insights → Platform usage in the newer admin interface). You can export a detailed CSV containing the user, feature, agent, timestamp, and credits consumed. Documentation

There are a couple of limitations worth knowing

• CSV exports cover up to 3 months or 1 million usage events, whichever comes first.
• Export download links expire after 24 hours.
• I haven't found a supported public API for retrieving the same granular credit data.
• The published Data Lake schemas don't currently include dedicated Rovo usage or credit tables.

Since you're already exporting the CSV and building custom reports, you could bring that data into Analytics if that's something you'd like to do. There’s no API for Rovo credit usage yet (vote on https://jira.atlassian.com/browse/AX-2000 ), so the export click would still be manual, but I wonder if everything after it can be scripted. For example:

1. Automate the file processing. A simple Google Apps Script or scheduled Python script could watch for the export email, download the CSV before the 24-hour link expires (assuming Atlassian's authentication allows it), and add the records to a central dataset.
2. Clean up and consolidate the data. Remove duplicates using Usage Id, then summarize consumption by day, user, app, feature, and agent. That gives us a much smaller, more useful dataset to report against.
3. Connect it to Analytics. If you're dealing with significant usage volumes, perhaps consider loading the exports into BigQuery, Snowflake, or another supported database instead. Otherwise, Google Sheets should work fine for the summarized data. I wouldn't store all the raw events in Sheets, though, since that could grow pretty quickly.
4. Add some context to the reporting. We could map the agent IDs back to their names and potentially join the Actor ID against the organization data in the Data Lake. Assuming those identifiers line up, that opens the door to reporting by team, not just individual users.
5. Build in a simple safeguard. Have the script flag when a fresh export hasn't been processed in 35 days, so we don't accidentally lose historical data because someone forgot the monthly export.

It's not quite a fully automated solution, but it gets us pretty close. The only recurring manual task would be initiating the export. Everything else, from processing and deduplication to refreshing the reports, could run in the background.

I also noticed that there are already two related public suggestions

• AI-1028: More detailed Rovo usage reporting - currently marked In Progress. This focuses on giving administrators better visibility into how individual users interact with Rovo.
• AI-1357: Extended Rovo credit reporting and API access - covers the 90-day reporting limitation, historical reporting, and the need for a public API.

Neither specifically commits Atlassian to exposing Rovo usage through the Data Lake.