Release notes 2026 07 14 — version 3.0.0
Summary
Section titled “Summary”BoomiSight 3.0 makes Confluence a more useful, page-aware view of your Boomi estate. The release adds Documentation Health, Architecture and API bylines, BoomiSight Specialist for Rovo, safe Process Property and Cross Reference views, and process-derived configuration dependencies. It also improves dashboard completeness, timestamps, pagination, freshness, retry behavior, security boundaries, and authoring clarity.
What changed
Section titled “What changed”Page-aware Documentation Health
Section titled “Page-aware Documentation Health”- New Boomi Architecture and Boomi APIs bylines summarize the Boomi context that is genuinely present on the current Confluence page. They can show relevant macros, visible Boomi links, API context, documentation health, freshness, and review signals without falling back to unrelated space-wide facts.
- Byline popups can refresh the current page context. This rereads the current Confluence page only; it does not scan the whole space, edit the page, or silently call Boomi.
- The dashboard now includes Documentation Health, where space admins can run an explicit bounded scan, review health and freshness, and open findings through valid Confluence page links.
- Documentation Health distinguishes the last completed scan from the time cached results were loaded. Reloading cached results does not claim that a new scan occurred.
- Missing page context remains unknown or not cached instead of being reported as zero or healthy.
Safer macro authoring and process context
Section titled “Safer macro authoring and process context”- BoomiSight macros now use consistent product icons and clearer, full-width configuration dialogs that open when a configurable macro is inserted.
- The new Custom Process Properties and XREF macro shows safe structural information for author-selected Process Property and Cross Reference components. Component names link to validated Boomi Build destinations, Process Property rows can show an allowed-value count, and XREF rows show column and row counts.
- Raw defaults, allowed values, password values, environment overrides, XREF rows and cells, source payloads, and internal identifiers remain hidden.
- Detailed Process Snapshot macros can opt in to Include configuration dependencies. When enabled, the snapshot shows Process Property and XREF relationships derived from the selected process’s current Boomi Build graph, including safe provenance and meaningful added, removed, and unchanged summaries.
- Existing Process Snapshot macros retain their previous presentation until an author deliberately enables configuration dependencies.
- Canonical Boomi Build links now route to the appropriate BoomiSight experience: processes open Process Snapshot, while Process Property and XREF links open the custom structural view. Malformed, external, lookalike, inconsistent, and unsupported links fail closed.
- Process Snapshot deployment rows once again show genuine Boomi branch labels when available.
BoomiSight Specialist for Rovo
Section titled “BoomiSight Specialist for Rovo”- BoomiSight Specialist can answer from safe cached BoomiSight facts for the trusted current page or current Confluence space.
- It can summarize Documentation Health; environments, runtimes, deployments, packaged components, APIs, Process Properties, and XREF structure; and verified process relationships.
- Named lookups search the applicable safe cache before the compact response window is applied, while broad questions use aggregates and relevant examples instead of returning hundreds of records.
- Current-page questions remain isolated from unrelated space-wide facts. Missing page association is reported as unknown or not cached, not as zero and not as permission to use tenant-wide counts.
- Process relationship refreshes require explicit confirmation and stay bounded to the trusted current-page process context. Passive answers remain cache-first.
- Answers preserve freshness, partial, stale, truncated, and unknown states and do not expose raw values, credentials, identifiers, source URLs, XML, or source payloads.
Clearer dashboards and operational views
Section titled “Clearer dashboards and operational views”- Environment Comparison and Documentation Health tables now include visible page information and Previous/Next controls.
- Environment Comparison reads the complete bounded active-deployment inventory before filtering, summaries, table pagination, or export. Deployments beyond the first API page are no longer omitted.
- The Execution Dashboard multi-environment filter no longer fails when its cache key contains more than one environment.
- Dashboard, macro, byline, and export timestamps use readable local formatting instead of raw ISO values. Missing or invalid timestamps keep stable fallback text.
- Recent Executions now displays readable execution times.
- Freshness and empty states are more precise. Healthy views avoid redundant state prose, while stale, partial, unavailable, in-progress, retry-later, truncated, and unknown states remain visible and actionable.
- Cached deployment results prefer genuine sanitized process names, avoiding generic unnamed labels when a trusted name is available.
Admin controls, security, and reliability
Section titled “Admin controls, security, and reliability”- Space admins can Clear cached data without deleting Boomi credentials or configuration.
- Space admins can Disconnect and purge space data only after entering the confirmation phrase. This removes the saved token, configuration, cached data, sync metadata, configured-space registry entry, and configured-space marker for that space.
- Sensitive actions derive the current Confluence space from trusted Forge context rather than caller-supplied routing identifiers.
- Central redaction boundaries protect credentials, authorization material, account and user identifiers, token-shaped values, and sensitive error details while preserving useful status and display labels.
- Transient Boomi throttling and upstream failures use bounded retry and backpressure. Long throttles, overlapping refreshes, and cooldowns preserve the last good cache and return a truthful retry-later, in-progress, partial, or stale state instead of starting duplicate work.
- Scheduled refreshes do not mark stale, partial, rate-limited, or duplicate work as a successful update.
Admin action required
Section titled “Admin action required”- A site admin must approve the BoomiSight 3.0.0 Marketplace upgrade and its Confluence space-write permission when Atlassian presents the upgrade.
- In every space that was configured before this release, open BoomiSight Configuration and save the existing complete configuration once. This creates the configured-space marker used by the new bylines and configured-space surfaces.
- Run Verify Connection & Access after saving to confirm that the stored Boomi credentials still have the intended access.
- Run Scan Space Pages from Documentation Health before expecting page findings. The scan is deliberately explicit and is not started by ordinary dashboard refreshes.
Authors who want process-derived Process Property and XREF relationships must edit an existing detailed Process Snapshot and enable Include configuration dependencies. Existing snapshots are not changed automatically.
Compatibility and upgrade notes
Section titled “Compatibility and upgrade notes”- BoomiSight 3.0.0 supports Confluence Cloud and customer-authorized Boomi Platform API access.
- No Boomi credential, configuration, or cached-data migration is required. Existing space-scoped data remains in Atlassian Forge storage.
- The permission set now uses Confluence space-write access for the configured-space marker and no longer requests Confluence page-write access. Page reads remain necessary for bylines and explicit Documentation Health scans.
- Backend egress remains limited to
https://api.boomi.com. - The release adds Forge modules for bylines, Rovo, the Process Property/XREF macro, and validated Boomi Build link routing. Atlassian may require site-admin approval before the new major version becomes active.
- BoomiSight Specialist availability depends on Rovo being available and enabled for the Atlassian site.
- Licensing, supported product, and uninstall behavior are otherwise unchanged.
Known Limitations
Section titled “Known Limitations”- Documentation Health is an explicit bounded scan, not a continuous or scheduled crawl. Results can become stale until a space admin scans again.
- BoomiSight Specialist answers from safe cached BoomiSight facts. Answers can be stale, partial, truncated, or unavailable when the underlying cache is incomplete; the agent does not treat the cache as a live authoritative Boomi inventory.
- Process-derived configuration dependencies describe the selected process’s current Boomi Build graph. They are not a historical reconstruction of an older deployed process version.
- The Custom Process Properties and XREF macro documents author-selected components; selecting a component does not prove that a process uses it.
- Raw Process Property values and XREF table contents remain intentionally hidden. Only safe structural facts and counts are shown.
- API Governance can still show partial enrichment when optional Boomi APIM data is unavailable to the configured identity.
- Validated link routing accepts canonical HTTPS Boomi Build links only. Unsupported or ambiguous Boomi links must be configured through a supported macro flow.
Validation
Section titled “Validation”- The promoted BoomiSight 3.0.0 production build completed successfully.
- Release changes passed the full automated app suite, focused privacy and boundary tests, lint, Forge manifest validation, and production-dependency security audit.
- Authenticated Confluence QA covered configured and unconfigured spaces, dashboard tabs, bylines, Documentation Health, Rovo page/space isolation, Process Property and XREF links, process relationships, positive allowed-value counts, meaningful data states, pagination, exports, refresh confirmation, and retry behavior.
- Production-scale regression data verified later-page deployment discovery, named lookups beyond the returned response window, bounded relationship indexing, truthful truncation metadata, and safe omission of raw values.