What's new this week

This section summarizes what was released this week.

Fast lane GA features

These features are generally available for all customers. Slow lane customers and tenants will have these features automatically enabled on November 2, 2026.

ai-sre: The AI-SRE help bot now retrieves product documentation through Cortex Search instead of the legacy HelpIndex. It continues to include customer-uploaded documentation, giving fast-lane users more relevant answers for help articles, OPAL reference, and how-to questions. This replaces a separate embedded search index with the same Cortex-powered documentation source used elsewhere in the product.
ai-sre: AI-SRE now supports Claude Opus 5.5 with high reasoning effort on the fast lane. This provides a stronger reasoning option for complex investigation and monitor-authoring workflows while keeping slow-lane customers on the existing model configuration until a later promotion.
visualization: Vega legend rendering now uses the optimized implementation on the fast lane. This improves responsiveness when working with charts that have large or complex legends.

Slow lane GA features

These features are enabled for all slow lane customers and tenants today. Fast-lane GA was 2026-09-09.

monitors: Monitor notification templates support formatTime and adjustTime helpers, so alert messages can render readable, timezone-adjusted timestamps.

Public preview and private preview features

features are available by invitation only. features are available by inquiry.

dashboards: The v2 dashboard read and definition-storage path is enabled for selected customers. This provides an early evaluation path for the v2 dashboard API and storage format while writes remain intentionally controlled.
charts: Chart-query BFF routing is enabled for selected customers, routing single-visualization chart data through the mainapp chart-query service.
apm: Uninstrumented service detection is available in private preview for selected customers. External dependencies such as third-party APIs and services without OpenTelemetry instrumentation now appear as first-class entities in APM, identified by their server.address: they show up as nodes with RED metrics on the global, dependency, and endpoint-scoped service maps, and time spent waiting on external calls appears in the performance breakdown chart, with grouping by operation and port. When a service slows down, engineers can tell at a glance whether the cause is their own code or an external dependency, without digging through span attributes in Trace Explorer.

Bug fixes

The following issues are fixed in this week's release:

ai: AI-SRE no longer logs tool results in dataset and dashboard tools, reducing accidental exposure of tool-returned content. (OB-66477)
ai: AI-SRE memory tool failures now explain why a memory operation could not complete. (OB-68263)
ai: AI-SRE monitor creation now reports monitor-write errors and rejects invalid updates instead of silently accepting a bad change. (OB-66200)
alerts: Alert search matching is now case-insensitive across the alert explorer and alerts API. (OB-68011)
alerts: Context filters can now address keys that are not bare identifiers, including keys with spaces, hyphens, or dots. (OB-67040)
dashboards: Dashboard conversion now handles legacy axes, stage layouts, parameter mappings, and placeholder dataset IDs more reliably. (OB-67012, OB-67013, OB-67014, OB-68305)
filterbar: The filter tour now pauses while the input picker is open, and selected fields persist when suggestions remount. (OB-67463, OB-68236)
ingest: Ingest errors now identify the ingest token that caused them, and route pipeline values are stored consistently as text. (OB-66430)
monitor-v2: AI-SRE monitor triage now preserves the monitor's last-modifier identity, and AI-SRE investigation controls behave correctly when stepping through rate-limit choices. (OB-67968)
service-explorer: Host CPU and memory infrastructure metrics now resolve correctly in Service Explorer. (OB-67822)
support: Saved Jira account handling, support drawer status filtering, and in-app comment attribution are more reliable. (OB-68347, OB-68030)
upconvert: Legacy dashboard conversion ignores obsolete parameter descriptions and preserves unreadable column indices and ordering. (OB-68304, OB-67013)

REST API changes

REST API surface changes for this release are published separately in the REST API Changelog, which covers version changes, breaking changes, deprecations, and migration.

Terraform provider

The following updates are available in the Observe Terraform provider.

v0.14.68 (2026-09-29)

New features

observe_dashboard: Support dashboards with schema version 2.
observe_ingest_routing: New resources for ingest routing.
observe_typed_datastream: New typed datastream support.
observe_monitor_v2: Object tags support.

Bug fixes

observe_dataset: Do not dry-run when the plan suppresses every change.
observe_snowflake_outbound_share: Recreate the resource when its name changes.

v0.14.69 (2026-10-01)

New features

observe_grant: Expose missing RBAC v2 roles, including oauth_token_creator.
Dataset labels are now resolved by ID through /v1/datasets.

v0.14.70 (2026-10-02)

New features

observe_default_dashboard_for_tag: New resource for setting default dashboards for tags. (OB-59521)

Bug fixes

observe_datastream: Guard datastream type changes with force_destroy and replace grants with their target.

Coming soon

The following GA features are already available for fast lane customers, and will be enabled for slow lane customers and tenants next week.

design-system: Stellar UI refresh with Snowflake-themed colors, the Inter font, redesigned toasts, and refreshed empty states.
docs-search: Cortex Search for in-product documentation.
monitors: Monitor list and stats REST APIs, including the paginated monitor resource shape.