Snowflake Cortex Support

Cortex Agent, Document Processing, and Search Batch Query usage now appear as individual line items in your Snowflake cost breakdown, rather than rolling up under generic compute. You can attribute AI and search workloads to the teams and products running them, compare spend across workload types, and see which Cortex features are driving cost growth. The new granularity flows through Explorer, Dimensions, and Anomaly Detection automatically, with no reconfiguration required.

CloudZero Labs

Administrators can now enroll an entire organization into a Labs feature in a single action, rather than enabling it user by user. Org-level enrollment is available from the Labs settings page and applies to every current and future member of the organization. You can still enroll individual users when you want to pilot a feature with a smaller group first.

Improved
  • Views: The active View is now visible and selectable directly in the Explorer top nav, reducing a 4–5 click configuration panel workflow to 1–2 clicks.
  • Cloud Connections: Ingest date bounds can now be set on a connection to control exactly which months are ingested during a migration, preventing duplicate cost line items when transitioning from one connection to another.
  • Optimize: The Recommendations listing now supports a second Group By dimension, letting you slice recommendations across two axes — such as service and account — in a single view.
  • Analytics: A new Scheduled Reports panel in the dashboard toolbar lets you view all your Looker-scheduled dashboard deliveries without leaving CloudZero.
  • Explorer: Shift+Tab keyboard navigation now works in filter and group-by dropdowns, letting you move backward through options without reaching for the mouse.
  • UX: The CloudZero application is now at app.cloudzero.com; the Classic Experience link has been updated to legacy.cloudzero.com to reflect the completed domain migration.
  • User Settings: Preference updates now apply atomically and validate each change individually, returning clear errors for unknown paths or mismatched value types instead of silently failing.
  • Datadog Connection: The most recently settled month now uses Datadog's finalized bill-of-record costs instead of estimates, so reported spend more accurately reflects your actual invoice.
  • GitHub Connection: Copilot costs now reflect GitHub's new AI Credits billing model (effective June 2026), and per-seat attribution is restored after GitHub's SKU naming change silently dropped seat-level data.
  • Optimize: The Delete EBS Snapshots Older Than 180 Days recommendation now uses fresher resource data and filters out snapshots with less than $1 in recent spend, reducing noise from low-impact items.
  • CloudZero MCP: The get_cost_data tool now returns a row count showing how many underlying billing records were aggregated into each result, making it easier to identify high-volume cost allocation hotspots.
  • Explorer: Container CPU and memory efficiency views load up to three times faster for large organizations, reducing wait time when exploring usage by cluster, namespace, or workload.
  • User Settings: Several user and organization preferences — including timezone, display name, Explorer sort order, and onboarding state — now persist server-side instead of falling back to browser storage.
  • Resource Details: The cost chart in the resource details panel now inherits the date range you have selected in Explorer, so the data shown always matches the time window you were already analyzing.
  • Cloud Connections: Degraded connections now surface the specific diagnostic message from the backend — such as missing required tables — instead of generic staleness copy, so you know exactly what to investigate.
  • Users: When SSO Groups manage role assignments for your organization, role editing in the Edit User drawer is now disabled with clear guidance, preventing manual changes that would be overridden by SSO.
  • Analytics: Scheduled report plans now surface the dashboard or view title alongside a direct link, making it easier to identify and navigate to the content behind each scheduled delivery.
Fixed
  • Explorer: The Last 3, 6, and 12 Months presets now end at the final day of the most recently completed month, matching Classic Explorer behavior and removing partial current-month data from rolling comparisons.
  • Explorer: Thirteen internal billing dimensions are now hidden from dimension pickers platform-wide, and five dimension labels — including Invoice ID and Payer Account — now display with correct spacing.
  • Explorer: Views configured with a cloud provider group-by now display cost data correctly after saving or reloading, for views affected by this configuration.
  • Optimize: Recommendations for resources with mixed-case names are no longer silently omitted from results; a case-sensitive matching issue caused certain resources to be excluded from the recommendation cache.
  • Optimize: RDS instance recommendations (Delete, Upgrade, Rightsize, and Migrate to Graviton) now correctly surface RDS instances, resolving a resource-matching issue that had caused these recommendations to return no results in RDS-heavy environments.
  • Explorer: The filter bar no longer overlaps the sticky toolbar when multiple filters wrap to additional rows, keeping controls and data always visible.
  • Optimize: Removing a group-by from the recommendations table no longer leaves column headers visible with blank cells; the table now correctly restores its default column layout.
  • Optimize: Filtered recommendation lists can now be exported when a search term is active; previously, the search filter caused export requests to fail with an error.
  • Explorer: Re-selecting a renamed favorite from Favorites or Recents now correctly preserves its custom name instead of reverting to the auto-generated label.
  • Datadog Connection: Re-authentication now correctly handles full URLs and sub-organization site addresses, restoring access for connections that were broken after the recent domain migration.
  • Views: A view containing special characters in its cost dimension no longer causes the entire organization's view list to fail; all other views load normally.
  • Databricks Connection: Resource tags from Databricks billing data are now correctly mapped to resource-level attribution, improving cost-allocation accuracy for Databricks workloads.
  • Optimize: Deleted EBS volumes now clear from the Inactive and Detached EBS Volumes recommendation within approximately two days, preventing resolved items from lingering in your recommendations list.
  • Cloud Connections: Errors encountered while updating a connection now return clear, actionable messages instead of generic server errors, making configuration problems easier to diagnose and resolve.
  • UX: Org-wide messages now remain visible until explicitly dismissed, then suppress for one day — ensuring important notices are never silently skipped when multiple tabs are open.
  • UX: Slack and Jira OAuth callbacks now redirect correctly to the integrations settings page, so connecting or reconnecting those integrations no longer lands on a 404 error.
  • Explorer: The Billing Entity cloud-provider dimension is now available in the dimension selector, restoring parity with Classic Explorer for organizations that use it in financial reporting.
  • Cloud Connections: OpenAI cost dimensions for GPT-5.4 and later models are now correctly parsed, resolving truncated model names and mislabeled Flex tier and regional endpoint usage.
  • Azure Connection: MCA, CSP, and MPA connections using non-USD billing currencies are no longer incorrectly rejected during validation; only EA connections require explicit currency configuration.