We have a specific use case that is giving us fits. Here's the gist:
- jira query with status transition and assignment history (assuming it's possible) of CRs
- analyze with codex to create a report which would contain, at a minimum Backlog Metrics
- Total "Active": the total number of CRs not in status "done", "rejected", or "in prod".
- Untouched: the number of CRs which have no comments and no changes in status
- Wait and Rejection Metrics (90 Day Rolling)
- 90 Day Average Wait Time: the average amount of time a CR currently in status "development backlog" or beyond spent in "waiting for customer" status over the last 90 days
- 90 Day Rolling Rejections: the total number of CRs moved to status "rejected" over the last 90 days
- 90 Day Rejection Reasons: of those CRs rejected over the last 90 days, how many of each "Rejection Reason" occurred
- Goals - Mean time to resolution (MTTR): across all CRs which are currently in status "development backlog" or "sprint backlog" or farther along in the workflow, what is the average amount of time they took to go from status "delivery backlog" to "done"
In the past I looked a couple of marketplace apps but they didn't really fit our use case with specificity. So my big question is: can Rovo or a Rovo agent pull this data that I can then feed into Codex/Claude Code/Gemini for analysis. I'm thinking if we could just get the raw history of the work item the AI could do the rest.