Data Protection and Security Toolkit 4.7.0 has been released

---

⚡ Event-triggered searches start without waiting

Previously, all search types shared a single scheduling pool. A long-running manual, scheduled, or full-instance search could occupy every available slot, and content-update searches triggered by Jira or Confluence activity had to wait their turn.

Content-update searches now run on a dedicated scheduler with its own capacity, separate from and additional to the main pool.

The practical effect: when someone edits an issue or a page, the resulting search starts right away, regardless of what else is running on the instance. This is most noticeable on busy instances where scheduled and full-instance searches run regularly.

---

📋 Rows per page is now a per-user setting

Page size was previously a single global setting shared by everyone with admin access.

It is now controlled per user. Each admin sets their own page size, and it applies consistently across list views.

The ignore list now respects that setting as well. It was previously fixed at a very small page size regardless of configuration, which made longer ignore lists tedious to work through.

---

🕓 Template history shows why each search ran

History entries now include the reason a search was triggered.

For searches triggered by content updates, history also shows the title of the Jira issue or Confluence page that caused the run.

This makes it much easier to understand activity on a template at a glance — particularly on instances where manual, scheduled, and event-triggered searches all run against the same template.

---

🔍 A new, privacy-conscious analytics foundation

Product analytics have moved to PostHog, with a model built specifically around how the app is actually used.

Three populations are kept strictly separate — administrators configuring the app, end users viewing announcements, and instance-level search activity — with each event associated to its site for aggregate reporting.

The privacy design matters here:

• Analytics events are proxied through our backend, so no analytics key is exposed in your browser.
• Error details and personal data are never sent to the analytics platform.
• Analytics remain fully opt-out via the existing setting, now labelled simply “Analytics”.

The result is a clearer picture of which features are genuinely useful, which directly informs what we prioritise next.

---

🛠️ Fixes and smaller improvements

Search reliability

• Content-update searches now work correctly on hybrid instances running both Connect and Forge components.
• Template search on Confluence now starts and stops reliably on all instance configurations.
• The app’s own changes to Confluence content no longer trigger further searches, preventing unnecessary repeat runs.
• Full-instance search no longer stops when it encounters Jira comments whose author has been deleted or deactivated.

Confluence

• Personal space keys containing characters such as @ . : - are now accepted instead of being rejected as invalid.
• Confluence administration screens now load reliably in all Forge configurations.

Notifications and editing

• Default SMTP email notifications now send correctly.
• The announcement editor now applies changes immediately for a noticeably snappier experience.