docs(changelog): add DefectDojo Pro 3.4.304 release notes

docs/content/releases/pro/changelog.md

@@ -20,6 +20,38 @@ For Open Source release notes, please see the [Releases page on GitHub](https://
 
 ## October 2026: v3.4
 
+### October 30, 2026: v3.4.304
+
+New features:
+* **(Connectors)** Ingestion Health has its own page, **Settings > Operations > Ingestion Health**, listing every source of Findings: each Upstream Connector, each scheduled Triage Engine rule, and each Asset and scan type imported through the API in the last 30 days. Each row shows the last run and its result, the baseline, when the next run is expected and any open anomalies. Three anomaly kinds join the zero-Findings check: **Volume Drop**, **Stale** and **Auth Failed**. Each source can have its own thresholds or be snoozed, the **Ingestion Health Warning** notification goes out once per anomaly, and the **Ingestion Health Status** dashboard widget counts the sources that need attention. Requires the **Ingestion Health** feature flag, off by default. See [Ingestion Health](/connectors/upstream/ingestion_health/#the-ingestion-health-page).
+
+Enhancements:
+* **(Connectors)** A Connector shows when its next scheduled Discover or Sync will run. The Operations view labels its Type, Status and Trigger tags in the reader's language, and a Connector save that fails because the connector service cannot be reached says so in plain words. See [Managing Operations](/connectors/upstream/manage_operations/#when-the-next-run-is-scheduled).
+* **(Assets)** The Open Finding Severity tile on the Asset page is one bar split by severity, with a legend across its full width and the total in the tile header, and its severity links filter correctly in every language. The Reachability Coverage panel leads with the Criticals confirmed reachable and the Criticals demoted, and shows coverage as one bar. See [Assets](/asset_modelling/engagements_tests/pro__assets/#asset-view) and [Reachability](/triage_findings/finding_scoring/reachability/#what-you-see).
+* **(Dashboards)** Each dashboard tile keeps its actions in one **Widget Actions** menu, so its controls no longer shrink or overlap beside the title and freshness stamp. **Add to Report** moved into the same menu. See [Customizable Dashboards](/metrics_reports/dashboards/custom-dashboards/#step-4-arrange-then-save).
+* **(Settings)** The Feature Flags page groups features by area, with an index beside the cards, an **X of Y Enabled** count, filters by state, and one-line descriptions that expand in place. See [Feature Flags](/admin/feature_flags/pro__feature_flags/#opening-the-feature-flags-page).
+* **(Parsers)** A [Generic Findings Import](/supported_tools/parsers/file/generic/#package-details) JSON finding accepts a `purl` for its dependency, or `purl_type` and `namespace` alongside `component_name`, so the dependency location carries the right package type.
+* **(Performance)** Reimporting large reports costs less per Finding, filtering the Findings list by mitigated date and requesting `?prefetch=endpoints` are faster, and updating or deleting a single Finding through the API no longer loads its related records one row at a time.
+
+Bug fixes:
+* **(Permissions)** The built-in Writer role can add files to Findings, Tests, Engagements and Assets again. A Permissions dialog opened before its page finishes loading lists only that record's members. See [Set Permissions in Pro](/admin/user_management/pro_permissions_overhaul/).
+* **(Deduplication)** After consolidating Findings, moving a tool back to **As the parser groups them** splits them on the next reimport, which reactivates each folded-away Finding with its locations. The first grouping change on an instance backfills existing Findings before it takes effect, so a reimport in between no longer closes Findings and opens copies. Reimporting an older Test whose Finding was folded into another keeps that Test's location active. On the Dedupe Pools page, a **Preview Re-run** is kept when **Add to Pool** finishes, a pool created during a refresh is listed, and a delete waits for the pool's members to load. See [Deduplication Tuning](/triage_findings/finding_deduplication/pro__deduplication_tuning/#changing-a-tools-grouping) and [Dedupe Pools](/triage_findings/finding_deduplication/pro__dedupe_pools/#adding-assets-and-previewing-first).
+* **(Importers)** **Reimport Scan** works on a Test whose scan type differs from its test type name, such as a GitLab SAST Report. With Locations enabled, a dependency too long to store is left off its Finding with a warning in the log, instead of failing the whole import. See [Tests](/asset_modelling/engagements_tests/pro__tests/#reimporting-scan-results-ui).
+* **(Findings)** Updating a Finding through the API returns its recalculated priority, and an update with `recalculate_hash_code` stores the recalculated hash and deduplicates on it. When a close is refused, **Close Finding** shows one message with the server's reason, and a refused note is shown under the **Note** field.
+* **(Tables)** The **On**, **Before**, **After** and **During** filters on date-only fields match the day the viewer picked, including for viewers in a timezone ahead of the server's. See [Customizing Tables](/navigation/pro__table_customization/#filtering-columns).
+* **(Licensing)** On an instance over its license limit, OAuth and OIDC sign-in works again, and only changes that add usage are blocked: new Findings, imports and reimports, new Assets, new users and new Connector operations. Revoking access, editing or deactivating users, and rotating tokens keep working.
+* **(SSO)** The SAML settings help says that the NameID picks the account a login lands on, not a mapped Username or Email. See [SAML Configuration](/admin/sso/pro__saml/#which-account-a-login-lands-on).
+* **(Navigation)** Every in-app walkthrough can be finished, starts on a record the user can act on, and describes what is on screen. A highlighted sidebar row stays in view when the sidebar footer grows. See [In-App Walkthroughs](/navigation/pro__walkthroughs/#task-walkthroughs).
+* **(UI)** A dashboard tile's default title follows the reader's language, the toggles on Priority Insights and the Asset hierarchy panel are translated, and translated column headers fit in the Regulatory Scope widget. Screen readers announce whether a popup menu is open. A switch whose change is refused, such as a policy's **Enabled** switch on **Settings > AI Governance Policies**, flips back and shows one message. A long unbroken error wraps inside its message instead of widening the page, and the PSIRT **Applies To** panel keeps keyboard focus after each change. See [Language and Localization](/navigation/language_and_localization/#what-is-translated-and-what-is-not).
+
+Security fixes:
+* **(Authentication)** A SAML logout message that cannot be read is refused with a 400 response. See [SAML Configuration](/admin/sso/pro__saml/).
+* **(Authorization)** The Risk Acceptance page acts only on notes that belong to that Risk Acceptance. See [Risk Acceptances](/triage_findings/findings_workflows/pro__risk_acceptance/).
+* **(Hardening)** Triage Engine Slack messages show Finding titles, Asset names and Card Facts as plain text. The Helm bootstrap scripts no longer print the admin password; they name the Secret that holds it and how to read it. The application images apply security updates to their bundled components. See [Send a Slack Message](/automation/triage_engine/node_reference/#send-a-slack-message) and [DefectDojo Pro Installation Guide (Kubernetes / Helm)](/get_started/pro/onprem/kubernetes/installing_on_kubernetes/).
+
+Behavior changes:
+* **(Assets)** Unmanaged Devices is now called Asset Discovery: the sidebar entry under **Attack Surface**, the page and its walkthrough, and the feature flag's title on the Feature Flags page. Menu search still finds the page by its old name. See [Asset Discovery](/asset_modelling/pro__unmanaged_devices/).
+
 ### October 29, 2026: v3.4.303
 
 New features: