Skip to main content

v1.5.0

What's Changed​

🔍 Troubleshoot​

  • feat: Expand Knowledge Graph with per-resource ontology refactoring, topology edges, and redesigned label and attribute filters
  • feat: Add interactive guided tours for Troubleshoot, LLM Analyser, and AI Gateway views
  • feat: Expand trace and log integration support for Pinot and Elasticsearch with custom index mappings and trace grouping
  • feat: Add diagnostic feedback for empty log queries caused by mistyped label values
  • fix: Enhance incident triage accuracy by deduplicating historical events, preventing seed re-fires, and improving target workload resolution
  • fix: Refine alert triage handling for partially saturated histograms and lagging root causes
  • fix: Improve Knowledge Graph search performance and ensure automatic cleanup of stale external service nodes
  • fix: Enhance Knowledge Graph node mapping accuracy for StatefulSet pods, integration instances, and UI icons
  • fix: Fix Knowledge Graph filter cascading and ensure workload criticality settings persist correctly
  • fix: Improve telemetry database stability by bounding connection pools, supporting null DB scan values, and isolating query block metric labels
  • fix: Correct eBPF network map source/destination labels and refine trace type-classification for consumer spans
  • fix: Resolve UI layout and menu selection issues in the investigation sidebar

⚡ Optimization​

  • feat: Enhance Optimize UX with a unified cost and reliability view, rich tooltips, blast-radius impact role categorization, and automatic rightsizing PR updates
  • feat: Introduce observed coverage tier allowing single-source tenants to safely reach impact bands
  • fix: Categorize requests-unset pod recommendations under Configuration and key rightsizing behavior on rule name
  • fix: Resolve recommendation count mismatches and optimize table performance on the Optimize page
  • fix: Automatically archive recommendations for inactive resources and close resolution cursors on scan errors

🔄 Workflow​

  • feat: Add centralized automation execution dashboard and standalone Task Runner page under Automations
  • feat: Restructure Automations page to tenant level with account filters and route under /automation
  • feat: Support workflow version pinning for Call Workflow actions, displaying last execution versions in automation listings
  • feat: Add interactive foreach sub-task editor and iteration execution views
  • fix: Add SSRF protection to runbook server network tasks and surface real automation builder save errors
  • fix: Surface inactive status reasons when attempting to trigger a workflow

☸️ Kubernetes​

  • feat: Surface service endpoint readiness directly in Kubernetes Services views
  • fix: Resolve StatefulSet-ordinal pod names to their owning workload in Knowledge Graph
  • fix: Prevent trace-observed Kubernetes nodes from being misclassified as Service
  • fix: Batch alert rule inserts and match count updates during Kubernetes collector discovery

☁️ Cloud​

  • fix: Acknowledge AWS EventBridge SQS messages when account tokens lack an active cloud account
  • fix: Gate Azure recommendation fallback logic strictly on the cloud provider
  • fix: Render GCP cloud logs using a standardized tabular CloudLogsTable view
  • fix: Skip AWS throttling during region fan-out to prevent alert flapping

🤖 AI​

  • feat: Add global Nubi AI chat drawer with a sticky launcher, nested task visualizations, and vision support for image attachments
  • feat: Add Google Vertex AI support on the LLM gateway with tenant bring-your-own-credentials
  • feat: Expand AI Gateway with Governance dashboards (user quotas, routing rules, model deprecation shields, DLP egress secret detection), cross-provider model substitution, and an OpenAI-compatible /v1 API endpoint
  • feat: Launch gateway Sessions dashboard for conversation rollups, search, message history, and audited body views
  • feat: Introduce Memory-v2 semantic RAG retrieval with citation attribution in AI responses and UI audit trails
  • feat: Add internal Critique Analytics dashboard, prompt-trace viewers, and LLM response cache hit rate KPI cards
  • feat: Update AI model catalog supporting OpenAI GPT-5.x, Claude, and Gemini 3.1 Flash-Lite models
  • feat: Expand AI agent capabilities with structured event investigation tools, RabbitMQ management access, and Cloud CLI fallbacks
  • fix: Recover from empty LLM responses via model fallbacks, enforce per-provider thinking-token spend limits, and resolve cross-provider config leaks
  • fix: Prevent agent re-decision loops by surfacing duplicate tool calls and routing automation creation to dedicated agents
  • fix: Improve agent tool execution reliability for log fetching, metrics querying, gcloud table formatting, and notebook turns
  • fix: Improve AI Chat UI performance by batching per-message feedback requests and ensuring Nubi chat renders properly over active modals
  • fix: Resolve @agent mention parsing issues with punctuation and prevent duplicate auto-PR creation and comment spam
  • fix: Allow read-only user roles to submit and view AI conversation feedback

🔔 Notifications​

  • feat: Redesign alert notifications with severity color stripes, hybrid Datadog-style cards, in-card action buttons, and automatic thread resolution replies in Slack
  • feat: Add cost posture alerts with accrued-waste context and SLO alert deep-links
  • fix: Surface Slack bot-membership errors clearly when adding channel members
  • fix: Silence cleared findings from Slack and render card facts as clean fields

🎫 Tickets​

  • fix: Enhance ticket creation and update event modals
  • fix: Restrict incident-task integration dropdown to supported ticket providers
  • fix: Roll back upgrade plan deletion when the target plan does not exist

🔌 Integrations​

  • feat: Introduce dynamic RBAC with user-defined custom roles and granular per-module permission controls
  • feat: Expand global page search with misspelling support and indexing for Automations, Optimizations, and Investigation tabs
  • feat: Expand Discord integration with full template parity and invite workflows
  • feat: Upgrade core database migration engine to Atlas
  • fix: Ensure integration status updates commit prior to audit log writing
  • fix: Enable direct in-cluster database connectivity without requiring a second terminal
  • fix: Resolve tenant branding issues and apply custom branding to guided-tour copy
  • fix: Add keyboard navigation support for product tours and 24-hour snooze option
  • fix: Recover Zenduty alert labels for accurate subject resolution
  • fix: Remove duplicate create buttons from Knowledge Base and account context empty states
  • fix: Prevent internal error detail leakage across integration endpoints
  • fix: Update AWS integration tile UI layout