ServiceNow Ticket Integration
ServiceNow ticketing is now generally available to every organization. Turn any Optimize recommendation, anomaly, or incident into a ServiceNow incident, one at a time or in bulk, and status changes in ServiceNow sync back to CloudZero automatically. Monitors can also create ServiceNow incidents, so alerts flow straight into your existing response workflow.
CloudZero MCP
You can now create, update, and delete manual insights from your AI tools through the CloudZero MCP server, including editing an insight's status, effort, and description. Insights generated automatically, such as anomalies and recommendations, stay read-only, and changes require full data access.
Improved
- Azure Connection: New connections detect the agreement type from Azure automatically, Enterprise Agreement connections without a configured currency adopt the billing currency Azure reports, and the agreement type is locked once data has been ingested, which prevents edits that would duplicate spend in your cost history. The connection detail panel now shows the Azure billing currency, and you can set or correct the Enterprise Agreement currency there.
- AI Signals: The Overview now combines cost and token usage into one full-width chart with a Cost and Tokens toggle, so you can compare both views by model without scrolling past duplicate charts.
- AI Signals: Clicking a bar, line, or band in the AI Explorer chart now applies a filter for that group, matching Explorer so you can drill into spend without using the table.
- Monitors: Typing @ in the email recipients field now suggests teammates from your organization, so you can add colleagues quickly while still entering any external address as free text.
- Monitors: The monitors table now shows Created By, Active Alerts, and Last Triggered, and the monitor panel lists active alerts, so you can see who owns a monitor and whether it is firing.
- Monitors: The Create and Edit Monitor drawer is now organized into collapsible sections, with a cost preview so you can confirm a monitor's scope before saving, and the Anomaly Detection section shows a chart that updates as you change the custom threshold.
- Explorer: Info popovers on Explorer results, the Usage Metrics toggle, and saved views and insights, plus several platform tables, now explain table interactions and link to documentation, making sorting, resizing, and pinning easier to discover.
- Monitors: An optional Work Item column in the Incidents table shows the ServiceNow or Jira ticket attached to each incident, and a ticket you create or unlink appears immediately, so you can see which incidents already have a ticket without reloading.
- Explorer: The comparison drawer now shows Compare-to presets in a single row and uses Start and End fields for custom ranges, matching the comparison popover so choosing a comparison period takes fewer steps.
- Monitors: The percentage significance threshold now measures against the same baseline as the dollar threshold, so both settings give consistent results when deciding which anomalies raise an alert. Existing monitors with a percentage threshold may now alert on different anomalies.
- Optimize: Cross-location read recommendations for Google Cloud Storage buckets now use the fixed title "Reduce Cross-Location Read Costs", so the full action is readable in the Recommendations table instead of being cut off by long bucket names.
- Monitors: Saving a monitor that duplicates an existing configuration now shows a clear message with a link to the existing monitor, so you can open it instead of retrying.
- Cloud Connections: Anthropic Enterprise billing data now carries your organization's UUID as a tag, so you can match the account ID in CloudZero to the ID Anthropic shows in its console.
Fixed
- Telemetry Streams: Deleting an organization's only telemetry stream now completes, so the stream is fully removed and its name can be reused.
- GitHub Connection: GitHub Actions cost is now attributed to each user who ran a workflow instead of a single user, so grouping by workflow and user shows accurate spend.
- Cloud Connections: OpenAI Enterprise Codex usage with no assigned user now ingests correctly, so a single unattributed record no longer blocks the month's usage and cost data.
- Cloud Connections: Error messages from Anthropic Enterprise and OpenAI Enterprise ingestion no longer include user names or emails, so personal details stay out of logs and status messages.
- Cloud Connections: OpenAI Enterprise workspaces billed in USD now show accurate cost instead of $0, and Codex and API usage are priced correctly for those workspaces.
- Databricks Connection: Genie usage now appears as separate, groupable fields such as surface, channel, and agent, so Genie Agents and Genie Code costs are attributed accurately.
- AI Signals: Supplemental context data now lands within the expected five-minute window instead of occasionally arriving up to 15 minutes late; no data was lost during these delays.
- Optimize: The cost graph on "Delete Inactive DynamoDB Table" recommendations now shows spend for only the recommended table, rather than all DynamoDB spend in the account, so the savings opportunity matches what you see.
- Cloud Connections: AI Telemetry connections now report each data delivery to connection health, so they no longer appear in an Error state while data is arriving as expected.
- Azure Connection: Costs billed in a non-USD currency are now labeled with their actual currency during ingest instead of USD.
- AWS Connection: Creating a connection now returns a clear error message when the link to your AWS account is refused, so you can see what went wrong and retry.
- Explorer: The cost type selector now always lists your role's default cost type, so you can switch back to it after choosing another cost type.
- Monitors: Forecasts now use each monitor's configured cost type instead of real cost, so monitors set to a different cost type, such as amortized, are forecast on the same basis they are evaluated on.

