Skip to main content

1.1.0

What's Changed​

🏕 Troubleshoot​

  • feat: Redesign the troubleshooting dashboard to support application-level grouping, summary widgets, drill-down interactions, and exclusion of configuration change events
  • feat: Support service and resource ownership mapping with a rule engine and owner management UI
  • feat: Support ingesting Kubernetes ConfigMaps and Secrets as first-class nodes in the Knowledge Graph
  • feat: Display live CPU and memory usage statistics in pod container resource views
  • feat: Add "Created By" and "Status" filtering options to the Tools management page
  • feat: Add a tenant-admin User filter to the Manual Investigations dashboard
  • feat: Support server-side OpenCost spend synchronization from central multi-tenant OpenCost deployment
  • feat: Added standalone cost-server to support upstream OpenCost allocations as a library
  • fix: Resolve issues with 0% CPU reporting on nodes, node conditions/taints handling, and Out-of-Memory (OOM) event enrichment
  • fix: Automatically enable default log provider for agent-wired integrations
  • fix: Add warnings in Helm installation notes and logs when publicly-known default secrets are detected
  • fix: Improve database handling and Kubernetes Resource Recommender (KRR) parsing in the collector

🏕 Notifications​

  • feat: Add audit logging for notification and ticketing integration settings
  • feat: Support Service Account authentication and bot space-bindings for Google Chat
  • feat: Add space-binding management interface for Google Chat integrations
  • feat: Enhance Google Chat incident handling with default spaces, automatic bot joining, and permission status checks
  • feat: Enable automatic space joining and incident-aware context in Google Chat
  • fix: Resolve issue where notification rules failed to save due to a renamed action
  • fix: Harden message delivery and migration handling for Google Chat integrations
  • fix: Resolve issues with thread replies in Slack notifications
  • fix: Improve resilience of Microsoft Teams and Slack OAuth installations against dropped session cookies
  • fix: Display connection card for unbound Google Chat spaces in the UI

🏕 Tickets​

  • feat: Expose key ticket metadata as top-level output fields in the Get Ticket action
  • fix: Restore authorization type during credential verification for GitHub App connection tests

🏕 Workflow​

  • feat: Enable zero-downtime vertical scaling using Kubernetes in-place pod resizing (v1.35+)
  • feat: Automatically link the originating workflow in Pull Requests generated by workflow executions
  • feat: Support per-version status tracking and draft lineage in the workflow editor
  • feat: Emit audit events for workflow automations
  • feat: Tag runbook automation conversations with an Automation source label
  • feat: Support integration identity synchronization for mapping local users to external accounts
  • feat: Support deleting individual versions of a workflow
  • feat: Add a Google Chat Join Space action to the workflow builder
  • feat: Enable smart-polling and status listing for triggered automations
  • feat: Redesign workflow version history rows using design-system components
  • feat: Add a "Go to Chat" shortcut link to workflow execution views for LLM-related tasks
  • fix: Resolve invalid UUID errors for system event consumers
  • fix: Restrict account administrators from modifying tenant-scoped configurations
  • fix: Require the approval type parameter in workflow approval actions
  • fix: Improve robustness of LLM JSON extraction and maintain stable data structures on parsing failures
  • fix: Add support for Anomaly event types in the workflow event-trigger dropdown
  • fix: Ensure workflow retries execute the pinned version instead of the live version
  • fix: Ensure the live version is executed for scheduled workflows, sub-workflows, and rollbacks
  • fix: Update modification metadata only when genuine edits are made to a workflow
  • fix: Enable tool-aware project labels for tickets
  • fix: Enable timezone-aware date formatting for tasks and triggers
  • fix: Fall back to the live workflow version during retries if the pinned version is unavailable
  • fix: Correctly match legacy webhook URLs when internal identifiers are missing

🏕 Cloud​

  • feat: Support executing CLI commands directly from the recommendation panel
  • feat: Add Google Cloud Platform (GCP) deployment and change enrichment sourced from Cloud Audit Logs
  • feat: Add detailed Google Cloud Run metrics with response code breakdowns
  • feat: Capture structured log attributes for Google Cloud Platform (GCP) cloud logs
  • feat: Add a per-service refresh button for cloud resource tables
  • feat: Introduce an optimization details side panel with automated interpretation insights
  • feat: Add impact-weighted FinOps scoring (v1.0)
  • fix: Parse Azure resource IDs case-insensitively during database apply commands
  • fix: Resolve execution safety issues in the cloud collector
  • fix: Paginate Azure Cost Management usage reports to prevent data truncation beyond 5,000 rows
  • fix: Enforce BigQuery billing access checks as a hard blocker during Google Cloud Platform (GCP) onboarding
  • fix: Decouple post-report synchronization from spend-fetching failures
  • fix: Correct AWS CLI syntax for applying right-sizing mitigations
  • fix: Capture diagnostic evidence for Google Cloud Platform (GCP) metric alert events
  • fix: Resolve issues with Google Cloud Platform (GCP) evidence loading, including deployment diff cards, the structured log viewer, and resource-driven log scoping

🏕 AI​

  • feat: Support per-conversation and per-category model selection, and resolve model tier inheritance issues
  • feat: Add triage heuristics and event-driven routing rules to the events investigation agent
  • feat: Add support for routing retrieval/summary agents to custom HuggingFace vLLM endpoints
  • feat: Enable automatic saving of AI-generated automations and improve automation builder chat stability
  • feat: Introduce AI Cost Analyzer featuring leaderboard analytics, usage reports, and optimization tips
  • feat: Introduce a user feedback tab in the LLM settings modal (available to super-admins and tenant-admins)
  • fix: Unify anomaly engine, reduce noise, and enable error-rate checks by default
  • fix: Make the "Open in Editor" link absolute so it renders correctly in Slack
  • fix: Default the workflow automation builder to the current active account and cluster context
  • fix: Resolve DBSCAN ValueError crashes caused by infinite or NaN values
  • fix: Prevent category-less sub-agents from inheriting parent tier reasoning credentials
  • fix: Surface specific rate-limit and quota-exhaustion details in AI chat rather than generic errors
  • fix: Clear stale validation alerts in the automation builder and migrate dropdown selectors to Design System components
  • fix: Ensure the final answer is persisted when a tool execution request is declined
  • fix: Resolve routing misdirection and duplicate saves in the workflow builder
  • fix: Prevent the trace tool from returning empty results on database query failures
  • fix: Improve shell command execution stability, input validation, and user feedback hints
  • fix: Improve workflow-builder determinism, block configuration leaks, and enhance validation build observability
  • fix: Configure the coding agent to clone the target branch of the Pull Request rather than the repository's default branch
  • fix: Enable intent-based routing for workflow builder turns to support unified edits
  • fix: Automatically retry HuggingFace API requests that return 408 timeout errors
  • fix: Cap inputs for asynchronous summarization and log retrieval to prevent token size bloat
  • fix: Correctly handle containers without resource definitions in cluster right-sizing fallback calculations
  • fix: Gate event analysis before evidence loading to prevent Out-Of-Memory (OOM) crash loops
  • fix: Display the correct "Last Refreshed" timestamp on optimization tabs

🏕 Integrations​

  • feat: Add support for Apache Pinot integrations, including auto-column dropdown selections and automated secret generation fixes