Display Density

• Access Controls: Roles with Cloud Integrations read or write permission can now view and manage billing Reviewable Periods, so the people who manage your cloud integrations can review and clear them directly.
• Access Controls: Required permissions on a role's permission screen now appear checked, with a lock icon and an explanatory tooltip, so you can see at a glance that they are already granted and cannot be changed.
• AI Signals: The AI allowance card now matches the Total cost card's month-to-date spend, including today in your time zone.
• AI Signals: The Overview month-end forecast now uses the same model as the Year-to-Date AI Spend report and shows a likely range, so forecast figures may differ from before. The pace and run-rate toggle is removed.
• Analytics: You can now select several dashboards and move them to a folder in one step, so reorganizing many dashboards no longer takes a separate dialog for each.
• API Keys: Keys with the Send AI Telemetry scope can now register an AI data plane and send AI usage data to the new telemetry ingest endpoint, so one key you already manage covers both enrollment and ingestion.
• Azure Connection: Agreement type detection now tells you what to fix during setup: it names the Azure reader role your tenant still needs to grant, or asks you to check the ID for typos when Azure can't find it, so you don't assign roles you don't need.
• Azure Connection: Azure connections now reuse their access token instead of requesting a new one for every call, which roughly halves the time for ingestion, connection validation, and agreement type detection.
• Cloud Connections: Billing periods awaiting review now appear on the Connections page, with Healthy, Error, and Reviewable Periods summary tiles, period details and actions, and rows that update as soon as you clear or silence a period, so you can act on billing issues without leaving the page.
• Cloud Connections: Adding a connection now follows one consistent flow for every supported provider and reports stalled requests sooner, and the connections list loads faster with each connection's health and the newest connections first, so you can spot setup problems and unhealthy connections quickly.
• Explorer: Month to Date and Year to Date daily ranges now end on the last complete day, so the final bar no longer looks like a drop in spend, and daily date headers in the cost table read "(partial)" for days still receiving data.
• Explorer: The export window in drilldown mode now separates Export Current View, which includes every expanded row and group-by column, from Export Summary, so you can see exactly what each file contains before you download it.
• Explorer: Comparison views now show the earlier period on the left and the current period on the right in charts, tooltips, the summary, and the cost table, and the inline comparison editor uses the same Start and End date fields as the comparison drawer, so you can set exact periods consistently.
• Collector: Settings and Setup now open in one shared window that comes to the front when requested, so moving between them is quicker and the tray app keeps running when you close it.
• Monitors: AI incident analysis now refreshes the title, summary, and description whenever an incident updates, including when its days change, so what you read always reflects current figures.
• Monitors: The Cards or List view you choose for alerts is now remembered across visits, so Monitors opens in your preferred layout without switching each time.
• Monitors: Service and Account Monitors now hide incidents whose cost impact is very small compared with the monitored service or account's typical daily spend, so you see fewer low-value incidents.
• Monitors: You can now delete monitors through the public API, so you can clean up monitors from your own automation. The incident details API no longer returns the undocumented filters and grouping fields; read a monitor's current grouping from its own details instead.
• Optimize: Administrators can now enable columns from user-created namespaces and Cloud Provider dimensions in Optimize Configuration, so recommendations can show the dimensions your team actually uses.
• Telemetry Streams: The streams page now lists only customer-managed streams and hides system-managed ones, so the list shows just the streams you manage.

- AI Signals: Organizations with only billing data now see a clear empty state on Overview, Explorer, and My Work instead of blank charts, and these pages load faster because they skip unavailable telemetry queries.

• AWS Connection: The connection details page now shows the correct External ID again, and Test Connection uses the ID your role actually accepts, so healthy connections no longer report a false failure.
• Azure Connection: Agreement type detection now returns a clear message within 20 seconds instead of a gateway timeout, so you can pick the agreement type or retry rather than wonder whether your ID is wrong.
• Cloud Connections: Deleting a CloudZero-managed connection now fully removes it, including its stored credentials.
• Explorer: Compare Periods Split view no longer shows crowded, overlapping totals above bars on daily data, so charts stay readable while tooltips still show every value.
• Explorer: Left and Right arrow keys now move the cursor when you type in filter and group-by search boxes, so you can correct a search without retyping it.
• Explorer: The calendar selection now matches the date range label, so the dates highlighted in the picker are the dates you see applied to your results.
• Explorer: The drilldown table and its download no longer show Cost of change while Compare is on, since that value measured a different period and could be misread as the compare result.
• Collector: Capture now resumes after you approve the CloudZero certificate outside the app, so usage tracking no longer stays stopped once the certificate is trusted.
• Monitors: Notifications are now correctly held back when the view they depend on has been deleted, so you no longer receive alerts that your minimum cost impact setting should have filtered out.
• Monitors: The incident activity log now shows the previous and new spend status, such as "changed from Ongoing to Recovered", so you can follow how an incident progressed.
• Monitors: Alert card sparklines now show days with no spend as zero dollars, so the line no longer implies continuous spend where there was none.
• Monitors: Incident analysis now completes for incidents involving large amounts of resource detail, so you get an explanation instead of a failed analysis.
• Optimize: Recommendation detail pages and activity logs now load promptly during busy periods instead of timing out.
• Telemetry Streams: Sending a large batch with many invalid records now returns a clear error listing the first 10 problems and the total count, instead of failing with a gateway error.
• Telemetry Streams: A deleted stream is now removed immediately, so it can no longer block a custom dimension that referenced it.
• Views: Saving a view with excluded dimensions now works in organizations whose dimension names contain unusual characters such as trailing spaces, so these saves no longer fail with an invalid request error.