Every Monday, a Jira admin team could spend an hour in the audit log, or none at all. Most weeks it spends none, until a workflow change surprises a team. What works better is a short review with the same questions every week, done in the same order.
I'm Marcin, founder of Radrly. We make ConfigRadar, a configuration change history app for Jira Cloud. Below is the agenda I suggest, first with only native Jira and then with the app. Use whichever half fits your site.
The agenda
Three questions, in this order:
- What changed in the last seven days, and how much of it matters?
- Was each important change expected? (Linked to a request, a release, or a known admin.)
- Is anything still open from last week?
Write the answers in one place: a Confluence page, a work item, a shared note. The written record is the point.
The native version
Minutes 0 to 5. Open Settings > System > Audit log and filter to the last seven days. Scan the categories that tend to matter most: permissions, workflows, fields, screens and space roles.
Minutes 5 to 15. Open the entries that look important. For each, note who made it and why. If the entry only says that an object was updated, open the object itself and compare it with what you remember.
Minutes 15 to 20. Record decisions. If you keep evidence, export the log as CSV.
What to know about the native route, from Atlassian's page on audit activities:
- The log can't be sorted.
- The CSV export holds up to 100,000 events and includes all events, even when the page is filtered.
- The Author field does not show the user for changes made in user management (users, groups, application access).
- Retention is set under Actions > Audit Log Settings. Check it once so last quarter is still there when an auditor asks.
- There is no severity, so a renamed status and a permission grant to everyone look equally loud.
None of this stops the review. It just makes minutes 5 to 15 longer.
The ConfigRadar version
ConfigRadar imports the audit log every hour (every six hours on Standard sites with up to 10 users), takes a daily snapshot of 14 configuration object types and shows each change with a before and after diff, a severity and, for risky changes, an impact explanation. Only Jira administrators can open it.
Minutes 0 to 5, with ConfigRadar. Add the ConfigRadar config risk gadget, titled Config risk this week, to the dashboard you already open on Mondays. It shows changes in the last seven days split by High, Medium and Low, plus up to five recent high severity changes. Only Jira administrators see the figures; anyone else gets a short message, so it is safe on a shared dashboard. The open alerts and drift figures are Advanced.
The Config risk this week gadget on a Jira dashboard. The open alerts and drift line is Advanced.
Minutes 5 to 15, with ConfigRadar. Open the timeline and filter by severity High, then Medium, over the last seven days. Each entry opens to its diff. Permission scheme diffs show grants by name, for example "Browse spaces: Space role Developers", marked Added or Removed. The impact rules flag things like Browse spaces granted to Anyone, a field made required, or a workflow transition removed.
ConfigRadar Overview: changes and high severity changes in the last 7 days, open alerts and monitored objects.
Minutes 15 to 20, with ConfigRadar. Open the weekly digest in the app to check that nothing from last week is still unexplained, write the decisions in your record, and export the week as CSV for evidence.
All of the above is in Standard.
Advanced adds three things that suit this ritual:
- Watchlist alerts: rules on the schemes, workflows, fields or spaces you care most about raise an alert when they change. In the Alerts view you acknowledge each one with an optional comment, which becomes the record of question 3.
- Periodic summary: a Jira work item created weekly on the weekday you choose (UTC, Monday by default), with counts by severity and object type and up to 20 top changes. Minimum severity is Medium by default. Anyone who can browse the summary space can read it, so pick that space with care.
- Baselines and drift: save a baseline after an approved review, and the next review starts from what has drifted since.
One limit worth knowing. ConfigRadar records and explains changes but never reverts them; it shows revert steps, and you make the change in Jira.
Listing: ConfigRadar on the Atlassian Marketplace.
Do you run a regular configuration review, and if so, which question takes your team the longest to answer?
