What Gets Collected
For each repository, Pawtograder stores two kinds of records:- Items — individual GitHub events with author, creation date, state, URL, and (where applicable) file-change summaries or labels. Item types are:
commitpr(pull request)issueissue_commentpr_review_comment
- Daily aggregates — per-day counts of each item type for the activity chart.
Enabling Analytics
Analytics are off by default. To turn them on:- Open the assignment configuration page (create or edit).
- Check Enable repository analytics.
- Save.
Refreshing Data
Two refresh paths exist:- Daily background refresh. A cron job runs hourly and, during the 2:00 AM UTC hour, enqueues one eligible assignment per minute (staggered to spread GitHub API load). An assignment is eligible if it is released, not archived, has repository analytics enabled, and has a due date on or after the current UTC date. Bulk fetches are batched in groups of 20 repositories and dispatched through a low-priority async queue.
- Manual refresh. The Refresh button on the dashboard queues a fetch for the current repository. The fetch status panel shows the last fetched timestamp and the current state (
idle,fetching,completed,error).
Manual refreshes are rate-limited to once per 10 minutes per repository. If you hit the limit, you are told when the next refresh is allowed.
Linking Rubric Checks to KPI Categories
Rubric checks can declare akpi_category that maps the check to one type of GitHub activity. When a check has a KPI category, the rubric sidebar shows a View [category] link that opens the Repo Analytics tab pre-filtered to the matching items.
Valid categories:
For the YAML syntax used to set
kpi_category on a check, see the Rubrics reference.