One browser. Your cluster estate. Smarter VM placement, capacity safeguards, real-time alerting, and clearer update triage—plus a first look at the vision for AI-assisted operations.
I am excited to announce version 1.0.0 of the Azure Local Cluster Management Tool!
This is a milestone I have been looking forward to sharing. The ambition is simple: make the everyday work of running Azure Local feel more connected, more visible, and less fragmented.
If you have ever moved between management consoles, remote sessions, and PowerShell windows just to understand what needs attention, you will recognise the problem. The tool brings cluster operations, performance, updates, alerting, and Azure Arc visibility together in a browser-based experience.
And there is another part of the story I am excited to introduce private AI integration, with SRE Agent as one integration path—not the limit of the vision. More on that below, including where I want to take this longer term.
Built from the ground up. Shaped by fleet-scale experience.
I built this tool from the ground up, drawing on my own experience managing hundreds of Azure Local clusters. It brings together deep Azure Local expertise and massive-scale fleet experience—a practical perspective on what operators need when one cluster becomes an entire estate.
That experience shapes the priorities: seeing where attention is needed across the fleet, getting to meaningful diagnostic detail, and keeping day-to-day workflows connected. This is not a management experience designed in isolation; it is built around the realities of operating Azure Local at scale.
Built around the work operators do.
The excitement around 1.0.0 is not about a version number. It is about bringing the operational picture together: what’s running, what is healthy, what needs updating, and where you need to look next.
The established management experience spans a single cluster or a wider estate, with capabilities that include:
- Fleet visibility: multi-cluster dashboards, fleet-wide VM and update status, favourites, and personal or shared cluster views.
- VM and node operations: start, stop, restart, and live-migrate virtual machines; inspect VM details; and conduct node pause/drain, resume, and failback operations.
- Performance and diagnostics: historical node and storage views, IOPS, latency, and throughput trends, alongside cluster events and remote log access.
- Solution updates: available updates, run history, live progress, and scheduled update windows across clusters.
- Operational control: threshold-based alerts, Teams and email delivery, maintenance windows, audit logging, and fine-grained access controls.
- Azure Arc visibility: registration, machines, extensions, Resource Bridge, custom locations, and Azure Local Sites.
That operational foundation continues in 1.0.0, with new capabilities focused on placement, capacity, alerting, and fleet triage. Explore the project overview for the broader feature set.
What is new in 1.0.0: sharper insights, stronger safeguards.
Update management deserves a spotlight. Knowing an update has failed is only the beginning; knowing which task failed is what helps you decide where to investigate.
The redesigned Fleet Update Status page puts triage first. A cleaner layout helps you focus on what needs attention, with optional advanced filters and an optional update-run-details panel when you need a deeper view.
VM placement and capacity: see the risks before you act.
The new VM Storage Alignment tab flags VMs whose compute placement does not match the Cluster Shared Volume hosting their disks. Capacity checks and preferred-owner context give you more information before deciding whether to change placement.
VM footprint tracking brings together configuration, checkpoints, disks, and a Save-VM estimate. A Save-VM CSV space warning and the CSV Save-VM Space Risk alert rule highlight storage pressure, while RAM-headroom warnings add capacity context to Live Migration and node Drain operations.
Node insights and configurable health thresholds
The opt-in Per-Host Process Insights view on the Nodes page shows the top CPU and memory processes, with a matching alert rule—helping you look beyond host-level utilisation to the processes behind it.
Configurable Health Service thresholds cover storage volume capacity warning/critical levels and minimum free node memory. Built-in safety checks flag values far from the defaults before they are applied.
Cluster health faults now identify their real source—live cluster data or Log Analytics—and are corroborated across sources when available. Quorum/witness degradation is explicitly visible too.
Real-time alerts where your operations team works.
Version 1.0.0 adds real-time direct alerting to Splunk and Microsoft Log Analytics, alongside the existing Teams and Email channels. Teams and Email delivery can each be disabled independently without losing their configuration.
This is distinct from the SIEM audit-log integration introduced in v0.13.0, which forwards audit—and optionally alert—activity to Splunk through HTTP Event Collector and/or Microsoft Sentinel through the Logs Ingestion API. That integration uses a pluggable design for future targets. Together, these capabilities broaden both operational alert delivery and audit visibility; they are not the same delivery path.
Less noise, smoother administration
- Maintenance-window retention: configure a retention policy and use the on-demand Run prune now admin action.
- Provisioning grace period: configure a grace period for new/un-clustered VM alerting to reduce noise while VMs are being provisioned.
- In-app update check: see a nav bar badge and an Admin > Settings panel for application updates.
- Cluster administration: the admin > Clusters Add/Edit form now opens as a modal dialog instead of an inline block.
- Versioned REST API routes: the new /Api/v1/… routes provide a versioned interface; existing routes continue to work as deprecated aliases.
Reliability fixes that matter during real operations.
The Disk Replacement Wizard fixes address the risk of targeting the wrong physical disk when several disks share the same drive model, a stale-cached-data edge case, and a PowerShell parameter error on the final Remove step.
The Live Update Monitor no longer reports a silent “complete” when no real output was produced. The misleading “System health OK” banner on the Solution Updates page has also been fixed for cases where the cluster’s update health state shows a real issue.
Further fixes address column-header tooltips missing in browsers with older cached assets and Settings dropdowns briefly showing stale values after a save. Splunk, Sentinel, and Log Analytics setup errors now have clearer troubleshooting guidance.
Stronger security boundaries for operators and automation
- Permission-safe action discovery: API responses are hardened so one caller’s permission level cannot be served from another caller’s cached result.
- Fairer API rate limiting limits partition by API key identity rather than relying solely on source IP, improving fairness for automation clients sharing network egress.
- Safer cluster registration: address validation rejects malformed URLs and unsafe targets, including loopback, unspecified, and multicast addresses.
- Validated integration endpoints: admin-configurable SIEM/Alerting URLs receive endpoint validation, with explicit warnings when TLS validation is disabled.
- Cleaner admin errors: messages no longer expose raw GUIDs or Azure resource paths.
That is the direction I want this tool to keep following: less hunting for context, and a clearer path from an estate-level overview to the detail that matters.
A look ahead: AI integration beyond SRE Agent
Alongside the 1.0.0 announcement, I am sharing a glimpse of a private AI integration feature. SRE Agent is one integration path, but the direction is not locked to SRE Agent: similar AI operations tools and agents could also be integrated.
The long-term goal is to evolve this into an Offline-First AI Operations Platform for Azure Local—with local operational context and AI-assisted investigation at its core, rather than making cloud connectivity the foundation of every AI workflow. That is the ambition, not a claim that the current release already delivers offline AI capabilities.
The AI integration remains private, not generally available. The broader vision is to support complementary tools and help operators investigate with better context—not replace engineering judgement. Compatibility with other agents, offline capabilities, and public availability will be shared as they are validated. This announcement does not promise autonomous remediation or a delivery date.
Version 1.0.0 is a milestone—not the finish line.
Explore the release—and help shape what comes next.
To explore the tool, start with the GitHub releases page and review the release-specific notes and assets. For deployment prerequisites and setup guidance, read the Quick Start guide before making changes to your environment.
If you are already using the tool, I would love to hear which workflows are helping you most—and where there’s still friction. Fleet visibility? Solution updates? Troubleshooting? Those are exactly the conversations worth having as the tool moves into its next chapter.
Hello, 1.0.0. Let us keep making Azure Local operations simpler.
No responses yet