Open any connected cluster and see what's actually running inside it right now, not yesterday's snapshot. Live visibility spans 21 Kubernetes resource pages with detail drawers and an overview of failing and stuck pods, and the same experience lands in ZopDay. Cost recommendations also gain a richer drawer that answers savings, confidence, effort, and what the resource is, all in one place.
New Live Kubernetes visibility across 21 resource pages (Pods, Deployments, StatefulSets, DaemonSets, Jobs, ConfigMaps, Secrets, Services, Ingresses, PVCs, Nodes, Events, and more) with type-specific detail drawers
New A cluster Overview that surfaces running and failed pods, healthy deployments, total restart count, and any pods stuck in crashloop, OOM, or pending, together with the warning events explaining why
New The same live cluster experience in ZopDay, plus a dedicated Settings page (node pools, autoscaling config, Kubernetes version)
Improved The recommendation drawer now answers four questions at a glance: monthly and annual savings, a confidence score, implementation complexity, and the resource's key specs, plus when it was first generated and last re-evaluated
Improved Cloud Estate -> Resources is now Inventory, with a new icon; routes and bookmarks are unchanged
ZopNight now derives the owner of each resource from cloud activity logs (CloudTrail, GCP Audit Logs, and Azure Activity Logs), so 'who created this?' is answerable in showback without digging through logs by hand. ZopDay adds in-place updates and cross-VPC connections for provisioned databases and clusters, and discovery now survives pod failures.
New ZopNight: resource ownership: the IAM principal that originally created each resource is derived from CloudTrail (AWS), Audit Logs (GCP), and Activity Logs (Azure), separating humans from service accounts and feeding owner / team accountability into showback
New ZopDay: change tier, storage, scaling, and backup retention on RDS, ElastiCache, EKS node groups, Cloud SQL, Memorystore, GKE node pools, Azure SQL / MySQL / Postgres, Azure Redis, and AKS in place, without recreating anything
Improved ZopDay: connect a datastore to a cluster across VPCs: ZopDay auto-creates and accepts VPC peering, adds reciprocal routes, and opens a narrow port-scoped ingress rule; cross-region connections are rejected up front with a clear message
Improved ZopDay: the Provisioning Jobs list now distinguishes CREATE / UPDATE / DELETE / CONNECT / DISCONNECT with colored badges and a Type filter, with human-readable step labels
Improved ZopNight: discovery work now survives pod failures: orphaned tasks are reclaimed and retried, so a restart mid-run no longer drops resources
ZopNight now discovers, prices, and generates cost recommendations for AWS Bedrock: your provisioned throughput, custom and imported models, knowledge bases, agents, and guardrails appear in topology with real cost attached, and ten new recommendation types flag idle and underutilized generative-AI resources. ZopDay also gets a guided first-run setup, and this release closes a class of invalid recommendations on orchestrator-managed compute.
New ZopNight: AWS Bedrock support: discover, price, and optimize Provisioned Throughput, custom and imported models, knowledge bases, agents, and guardrails, with 10 new recommendation types (idle throughput, unused custom models, old model generations, and more)
New ZopDay: a guided three-step first-run setup: pick your goals, connect a cloud account, and choose where to start (deploy a service, provision infrastructure, or connect a datastore)
Improved ZopDay: clear connect-an-account prompts on Infrastructure, Datastores, K8s, and Deploy when no provisioning account is connected, instead of empty lists or broken flows
Improved ZopNight: keyboard navigation on the Architecture Globe and Canvas, and collapsible per-platform setup guides when adding Slack, Microsoft Teams, Google Chat, or webhook notification channels
ZopDay gets one-click service networking and a rebuilt Deployment Spaces experience. Configure how a deployed service is exposed (through the cluster ingress controller or a dedicated cloud load balancer) straight from the canvas, and navigate a cleaner, health-first Deployment Spaces layout that leads with what's running and what's connected.
New ZopDay: a new Network tab on the service drawer: route through the cluster's ingress controller (with optional cert-manager TLS) or a dedicated cloud load balancer, without editing Helm values or rerunning the create flow
Improved ZopDay: Deployment Spaces rebuilt around health first: a searchable list with status filters, a live provisioning stripe, a single connection-state status badge, and a Health Overview strip (nodes, CPU, memory, components installed)
Improved ZopDay: Platform Components grouped into All / Core / Optional / Needs attention, with a promoted one-click Install All Core action
Improved ZopDay: sidebar regrouped into Infrastructure, Deployment, and Activity, with a consistent Provision New action on every page
Improved ZopDay: Datastores & Caches now surface access errors clearly instead of failing silently
ZopNight is generally available. Connect AWS, GCP, and Azure in under a minute and start cutting spend across your whole footprint: schedule non-production resources off nights and weekends, and trim production with 337+ evidence-backed audit rules. No agents, no cluster deploys, read-only by default, and ZopNight only ever starts and stops resources.
New Schedule anything with a visual weekly grid: click and drag across a 7-day x time-slot matrix, no cron syntax required
New Dependency-aware resource groups start and stop resources in the right order: database before app server, and the reverse on teardown
New One-time overrides (force_on / force_off) with an expiry, for maintenance windows and incidents
New Connect AWS, GCP, and Azure in under 60 seconds: every mainstream auth method per cloud, with the exact IAM / role JSON generated live for your chosen permission tier
New Full inventory discovery with topology, 14-day metrics (CPU, network, disk, memory), and bulk start / stop
New Architecture view: a 3D globe of every region plus a full hierarchy canvas with inline recommendations and metrics on every node
New 337+ audit rules across AWS, GCP, and Azure surface idle, right-sizing, orphan, commitment, and governance savings with 14-day evidence and current-vs-optimized cost
New Auto-tagging predicts prod / staging / dev with confidence scoring and human approval, feeding showback attribution
New Scheduled VM autoscaling for ASG, VMSS, and MIG, with Business Hours, Peak Hours, and Weekend Scale-Down presets
New Event Readiness: pre-scale for traffic spikes with a guided wizard and automatic rollback after the event
New Dashboard, cost forecasting, showback by team and tag, multi-currency, budget tracking, and cost anomaly detection with root-cause breakdown
New Organizations, teams, and role-based access control with full audit logging and Slack, Microsoft Teams, Email, and Google Chat notifications
New MCP server with read-only tools for Cursor, Claude Code, Codex, and Windsurf
New API-first platform: 125+ REST endpoints with consistent pagination and personal access tokens
New Industry Playground: an instant, no-auth demo across finance, retail, and healthcare scenarios
ZopNight MCP Server is now generally available. Any MCP-compatible AI client, Claude Desktop, Cursor, Codex, can now query your cloud state, surface findings, and apply remediations with approval. This release also ships a significant improvement to anomaly detection sensitivity and expands GCP region support.
New MCP Server GA: works with Claude Desktop, Cursor, Codex, and any MCP-compatible client
New MCP tool: list_findings, query open findings with filters by account, impact, region
New MCP tool: apply_remediation, create pending approval records from AI chat
New MCP tool: get_showback_report, department-level cost data accessible from AI clients
Recurring schedules now support full cron expressions with timezone awareness. Slack notification integration ships alongside this release, giving teams visibility into scheduled actions without logging into ZopNight. Azure Scale Set support rounds out the compute coverage for Azure customers.
New Recurring schedules: full cron expression support with per-schedule timezone configuration
New Slack notifications: receive a message when a schedule executes, pauses, or fails
New Azure Scale Set support: schedules can now target Azure Virtual Machine Scale Sets
Improved Schedule execution logs are now retained for 90 days (previously 14)
Fixed Schedules no longer execute during the account connection warm-up window
The new Showback dashboard gives FinOps and finance teams a department-level cost view without requiring engineering support to build custom reports. CSV export and custom date range selection make it suitable for monthly reporting workflows.
New Department-level showback view: filter and group by team tag, account, or organizational unit
New CSV export: download showback data for any date range, compatible with Excel and Google Sheets
New Custom date ranges: select any 1–90 day window for showback reporting
Improved Showback data now updates every 4 hours (previously daily)
Fixed Showback costs no longer double-count shared Reserved Instance capacity
The recommendations engine now uses a 90-day utilization baseline, up from 14 days, giving much more stable rightsizing signals. Every recommendation ships with a confidence score and estimated payback timeline. One-click rightsizing for EC2 and RDS is now available for all accounts with sufficient permissions.
New 90-day utilization baseline for all rightsizing recommendations (previously 14-day)
New Confidence scoring: High/Medium/Low confidence ratings on every recommendation
New One-click rightsizing: apply EC2 and RDS instance type changes directly from ZopNight
Improved Recommendations now respect application-tier tags to avoid downsizing production services
Fixed RDS recommendations no longer suggest instance families unavailable in the target region
AWS Organizations support is now generally available, allowing customers to connect and manage up to 500 member accounts from a single ZopNight workspace. Cross-account RBAC lets you assign view or manage permissions to specific accounts by team or role.
New AWS Organizations support: connect all member accounts from a single management account
New Up to 500 member accounts supported per ZopNight workspace (Enterprise plan)
New Cross-account RBAC: assign per-account view or manage permissions to team members
Improved Account connection now uses AWS IAM Identity Center SSO where available
Fixed Atlas inventory no longer duplicates resources in accounts with cross-account trust relationships
Tag policies let admins gate showback views, schedules, and remediation approvals on resource tags. A new policy editor walks teams through scoping access to a department or environment without touching IAM. Tag inheritance from AWS Organizations is honored throughout the policy engine.
New Tag policy editor: gate ZopNight views and actions on resource tag values
New AWS Organizations tag inheritance is now respected in policy evaluation
Improved Policy violations now include a one-click "request access" flow for end users
Fixed Policy preview no longer mis-classifies resources missing the tag key
Drift Detection compares the live state of your AWS accounts against the last known good baseline and flags resources whose configuration has changed outside of scheduled actions. Email digests summarize drift overnight; the dashboard lets you accept or reject changes per resource.
New Daily drift digest emailed to workspace admins at 08:00 local time
New Per-resource accept / reject controls for drift events
New Slack notification channel for high-impact drift (security groups, IAM)
Improved Baseline snapshots now retained for 180 days (previously 30)
Get release notes in your inbox.
We ship every two weeks. Subscribe to receive a short summary of what's new, no marketing, just the changelog in email form.