The ZopNight sidebar was a long list of seven sections, all open at once. It now opens with Home, then the pages you use most under Pinned.
Improved ZopNight: a new sidebar, with sections you can fold. Pin any page, remove a pin, drag pins into the order you want, or reset them to the defaults of Inventory, Schedules, Savings Recommendations, and Cost Reports
Invoices now live inside ZopNight. Settings → Billing brings together your plan, every invoice issued, and your payment history. Anything outstanding can be paid on the spot through a secure provider, so your card details are never handled or stored.
New ZopNight: see and pay your invoices without leaving the app. Each one opens as a full document you can print or download; credit notes and refunds sit alongside the invoices they relate to, and your company and tax details are editable in the same place
New ZopNight & ZopDay: one sign-in page for every product, where you sign in and pick zopnight, zopcloud, or zopday. Your last-used product is remembered and opens directly next time. The existing ZopNight and ZopDay sign-in pages are unchanged
New ZopNight & ZopDay: a product switcher in the sidebar, so you can move between products without going back to a sign-in page. Switching to zopcloud happens in place, since it is a view of ZopDay rather than a separate app, and it now carries its own green mark
Improved ZopNight: more detail on your Copilot seats. A failed licence assignment now carries Microsoft's own error wording with a plain-English gloss, so a count violation reads differently from a dependency violation. Licences assigned through a group are marked as such, and seats below your activity threshold are now counted, so the figures reconcile against your total seat count. Each seat shows the days it was active, and how long since it was last used on each Copilot surface
ZopDay can now send a VM space's metrics to a store you own: VictoriaMetrics, Grafana Mimir, or Prometheus. The service's Observability tab reads them back in a new Usage section, charting CPU, memory, network, and disk activity over the last 15 minutes to 24 hours.
New ZopDay: send a VM space's metrics to a store you own. Connect VictoriaMetrics, Grafana Mimir, or Prometheus under Settings → Integrations, attach the backend to a space, and install the agent. The Observability tab then gains a Usage section, charting CPU, memory, network, and disk activity over the last 15 minutes to 24 hours. Backends are optional, and attached per space. Services on ZopCloud get the same tab. Their shared space carries the backend, and each customer sees only their own containers
New ZopDay: pod consolidation for Red Hat OpenShift (ROSA) clusters. Enable it under Automatic Scaling on the cluster's Settings tab, and pods are packed onto fewer nodes so the autoscaler can remove the empty ones. The first time you switch it on, ZopDay installs the add-on through the cluster's own operator catalogue. It also warns you when the cluster's placement rules would undo the packing
Microsoft 365 Copilot is sold per person, so the number that matters is not how many licences you bought but how many get used. Connect your tenant and ZopNight tracks every seat from purchase to last use, with the renewal date and what the unused ones cost you each month.
New ZopNight: Microsoft 365 Copilot seat licences. Connect your tenant on the Integrations page in Settings, and AI Spend in Cost Reports gains a Seat licences view. It shows how many seats you bought, how many you handed out and how many are actually used, when each person last used Copilot, adoption over time, the renewal date, and what the unused seats cost you each month
Improved ZopNight: recommendations now open with the fix, not the fault. Every title names the action to take, so you can work down a list at a glance. Inside, what we observed is kept apart from what we recommend; severity carries a sentence explaining its rating rather than just a colour; confidence reads High, Medium or Low instead of a percentage; and the resource's own details sit in one panel at the top
Improved ZopDay: AI assistants now go beyond deploys, builds and revisions. They can register and remove infrastructure, create namespaces, check a custom domain's DNS, install and upgrade components, run and retry database migrations, read build and service logs, and diagnose a failing deploy. That takes the catalogue from 289 tools to 321. Changes still need the same approvals as before, and running a database migration now requires a permission of its own
Connecting an AI assistant to ZopNight or ZopDay no longer has to mean copying an access token by hand. Paste the MCP URL once, approve the consent screen, and grant the assistant access to specific organisations, choosing what it can do.
New ZopNight: connect any AI assistant over OAuth. Paste the MCP URL into your assistant, approve the consent screen, and grant it access to specific organisations, choosing what it can do. It works with any assistant that supports custom MCP servers, including Claude Code, Cursor, Codex, Devin, and Grok. Connections are listed under Developer Settings, where you can revoke access at any time without affecting your other integrations
Improved ZopDay: a failed deploy on a shared VM no longer blocks other services on that machine. It cleans up after itself and fails on its own
ZopNight now keeps a full history of every recommendation, and re-reads the resource to confirm the outcome rather than treating a finding going quiet as proof it was fixed. Savings are claimed only where there is evidence.
New ZopNight: recommendation history, and confirmation of whether the problem was actually fixed. Click History on any recommendation to see when it was raised, viewed, applied, closed, and why
New ZopNight: bookmark the recommendations you are working through, so you can come back to them
Improved ZopNight: tags now reach the cloud on fifteen more AWS resource types, taking the total from 21 to 36. Apply covers CloudWatch log groups and alarms, ElastiCache, EFS, KMS keys, Glue jobs, web application firewalls, ACM certificates, Route 53 zones, API Gateway, OpenSearch, MSK, and Redshift, all of which were detection-only before
Improved ZopNight: AI assistants can now manage infrastructure-as-code governance. Fourteen new tools cover the Terraform and OpenTofu pull-request gates, per-repository scanning and its status, scan tokens, and validation runs with their audit trail among others, taking the catalogue from 275 tools to 289
Improved ZopNight: you can now choose any write capability when creating a Personal Access Token, whatever your role or your organisation's MCP write tier. Every option is available up front, and your role in each organisation is enforced when the token is used
Red Hat OpenShift clusters now appear in ZopDay alongside EKS, GKE, and AKS, and you can configure them. Machine pools can be resized by hand or handed to the cluster autoscaler. The MCP tool surface more than doubles, from 122 tools to 275. Revealing an API token now hands you a ready-to-copy config block for your AI client.
New ZopDay: Red Hat OpenShift (ROSA) cluster configuration. Clusters are discovered through OpenShift Cluster Manager and listed on the infrastructure page alongside EKS, GKE, and AKS. Resize a machine pool by hand, or let the cluster autoscaler do it within limits you set
Improved ZopNight: the MCP tool surface grows from 122 tools to 275, so agents can now manage schedules, budgets, dashboards, policies, teams, and recommendations among others, and start or stop resources. Revealing an API token in Developer Settings also gives you a ready-to-copy config block for Cursor, Claude Code, Codex, Devin, Antigravity, or Grok
ZopNight now discovers Red Hat OpenShift clusters on AWS and gives them the same live Kubernetes view as the rest of your estate, alongside OpenShift-native views. A new IaC Governance area checks every Terraform and OpenTofu pull request against policies you attach to a repository, posting a pass or fail on the PR without touching your infrastructure.
New ZopNight: Red Hat OpenShift (ROSA) support. Discover OpenShift clusters on AWS and get the same live Kubernetes view as the rest of your estate, plus native views for Routes, Deployment Configs, Build Configs, Image Streams, and Projects. Connecting a cluster is explicit and revocable. Red Hat manages the worker nodes, so these clusters have no start, stop, or scheduling controls
New ZopNight: IaC Governance. Attach policies to a repository, and every Terraform or OpenTofu pull request is checked against them, with a pass or fail posted on the PR. Each finding names the severity, the resource, and the fix. Mandatory policies block the merge; advisory ones only flag. Start from a built-in catalogue of cost, security, and best-practice rules, or write your own. ZopNight only reports; it never changes your infrastructure
Improved ZopNight: deleting or rotating an AI Gateway key no longer drops its historical spend from your cost reports
ZopNight now lets you write your own recommendation policies. Pick the resources and a metric, set a threshold over a time window, and choose what ZopNight should raise when that threshold is crossed. Custom recommendations appear beside the built-in ones and behave the same way, each explaining why it was raised and what to do about it.
New ZopNight: custom recommendation policies. Pick the resources and a metric such as CPU, set a threshold over a time window, and choose what ZopNight should raise when that threshold is crossed: Idle, Schedule, Rightsize, Retire, or a category you name yourself
New ZopNight: custom recommendations sit beside the built-in ones and work the same way. Each explains why it fired, estimates the savings, and warns before you stop something critical such as a database. Rules also catch short spikes, so a brief burst above a threshold like 80% CPU no longer slips through
Tagging policies no longer have to be built one rule at a time. Download the template, fill it in, upload it, and apply the whole sheet at once. Approved tags can now be written onto the real AWS, Google Cloud, or Azure resource rather than only recorded in ZopNight, and audit logs now capture the read-only lookups AI assistants make over MCP.
New ZopNight: build tagging policies in a spreadsheet. Download the template, fill it in, upload it, and review every rule ZopNight read back before applying the whole sheet at once. Existing policies export the same way
New ZopNight: applying an approved Smart Tag now writes it straight onto the real AWS, Google Cloud, or Azure resource, not just records it inside ZopNight. Existing cloud tags are kept and nothing is ever removed, and every accept, revoke, and apply asks for confirmation first, stating how many tags will go onto how many resources
New ZopDay: deploy straight from a public repository or a container image using a shareable link, choosing the branch from a searchable list
Improved ZopNight: audit logs now record the read-only lookups AI assistants make over MCP, shown as clearly labelled Read entries. A new source filter narrows the log to assistant activity, and repeated identical lookups collapse into a single line. Expect a busier log than before
Services on a VM no longer have to run around the clock. With sleep enabled for a space, an idle service stops on its own, frees its memory, and restarts the moment someone visits. Alerts now name the organisation they concern and open it directly, and schedule recommendations show exactly what applying them will do.
New ZopDay: sleep for idle VM services. A service on a VM can stop when nobody uses it, free its memory, and restart on the next request, with an idle window of 1 minute to 24 hours set per space in Infrastructure settings. Rolling out per VM pool; existing services keep running until it is enabled
New ZopDay: shareable app install links. Send a link that opens ZopDay in Starter Templates with an app already selected, so a WordPress install skips the manual search. New users land back on the same app after signing in
Improved ZopNight: a cloud account with invalid or expired credentials now says so plainly instead of a stale mix of permission states. If the check cannot complete, you see 'Couldn't check permissions' rather than a false pass or fail
Improved ZopNight: alert emails and chat notifications now name the organisation they concern, and clicking through opens that organisation directly instead of the one you last had open
Improved ZopNight: schedule recommendations now show what will actually happen. The editor prefills from the grid you just reviewed, the review step names the resource the schedule attaches to, and resources with stale metrics still show their history, labeled with the last data received
A single Alerts page now gathers everything that needs your attention across your cloud accounts, from expired credentials to provider incidents. When one underlying problem affects several services, you see a single alert instead of a copy for each. Cost recommendations are now ranked by how much you stand to save.
New ZopNight: Alerts. A single page for everything that needs your attention across your cloud accounts: expired credentials, missing permissions, quota limits, billing-sync failures, and provider incidents. One underlying problem raises one alert instead of a copy for each affected service, and resolving it clears the alert everywhere
Improved ZopNight: cost recommendations are now rated Critical, High, Medium, or Low according to how much you stand to save, measured against your own cloud spend. A larger saving never ranks below a smaller one. Security, compliance, reliability, and other non-cost recommendations are unaffected
Improved ZopDay: one-click chart upgrades. When a newer version of a chart is available for an imported Helm app or an App Catalog service, the Overview tab shows an update card. Choosing Upgrade lists what the new version adds and removes, then redeploys with your existing settings unchanged
Improved ZopNight: the schedule recommendation panel now shows the evidence behind each suggestion, including the hours a resource actually runs each week and a savings estimate that reflects exactly what applying the schedule will do
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.
Multi-cloud automation·Production-ready in 30 min·SOC 2 · ISO 27001·20–60% off the bill, first month·4 platforms · 1 console·Multi-cloud automation·Production-ready in 30 min·SOC 2 · ISO 27001·20–60% off the bill, first month·4 platforms · 1 console·