NeuBird
LoginDemo

NeuBird for VMware vCenter

VMware vCenter root cause analysis, from the app down to the host.

NeuBird is the AI SRE. It reads your vCenter inventory, alarms, events, performance metrics, host logs, snapshots and certificates alongside your application telemetry and Kubernetes, and traces an incident from the service that slowed to the ESXi host or datastore underneath it. It isolates root cause in under 5 minutes, at 94% accuracy.

The problem

Why does an application alert start in vSphere?The cause often sits under the VM.

The page says checkout-svc is slow. Nothing in the service changed. Underneath it, an ESXi host raised a hardware alarm and went into maintenance mode, and DRS moved its virtual machines, including two Kubernetes nodes, onto a host that was already busy. The application team sees latency, the vSphere team sees a routine evacuation, and neither view shows the other.

What NeuBird reads

What does NeuBird read in vCenter?Every metric, next to the event that explains it.

A CPU ready spike on its own is a symptom. NeuBird reads the vCenter events and tasks for the same minutes, so the spike on k8s-prod-w07 lines up with the DRS migration that put it on esx-dc1-12, and the cluster's vCPU to core ratio shows why that host could not absorb it.

  • Inventory

    Datacenters, clusters, ESXi hosts, VMs, datastores and resource pools.

  • Alarms

    Triggered alarms on every object, such as host hardware sensor status.

  • Events and tasks

    DRS migrations, vMotion, HA restarts, maintenance mode and who started each task.

  • Performance

    CPU ready, memory ballooning and swapping, datastore read and write latency.

  • Host logs

    vmkernel.log and hostd.log from each ESXi host.

  • Snapshots

    Every snapshot with its age and size, and VMs that need consolidation.

  • Certificates

    vCenter and ESXi certificates, including VMCA-issued ones, with expiry dates.

  • Queried in place

    No bulk telemetry stored. NeuBird asks vCenter when an investigation needs it.

Resolve

How does NeuBird trace an incident from the app to the host?One investigation, across every layer.

When the alert fires, NeuBird queries 15+ sources in parallel: application metrics and traces, Kubernetes pod placement, the vCenter VM behind each node, the host it runs on, the cluster's events and the host logs. It follows the evidence down until it reaches a cause, and posts the root cause with every source cited, in under 5 minutes at 94% accuracy. The fix is staged for your team to approve, and the conclusion is saved to memory, with no bulk telemetry stored, so the next host evacuation is recognized at once.

Prevent

Can NeuBird catch a vSphere problem before the page?Risks flagged 30 to 60 minutes before impact.

A datastore alarm says a threshold was crossed. NeuBird works out when the datastore will actually fill and what is filling it. Here, a snapshot delta on checkout-db-02 that vCenter marks as needing consolidation grows 6 GB a minute, and the datastore fills in 48 minutes. NeuBird stages the move before VMs on that datastore pause, and it watches snapshot growth, certificate expiry and host hardware alarms the same way.

Resolve, Prevent, Operate

What does NeuBird find in a vSphere estate?Before, during and between incidents.

Between incidents, NeuBird runs the hygiene: old snapshots, alarms that never clear, hosts that drift from the cluster's ESXi build and clusters left unbalanced after maintenance.

Resolve

Latency after a host evacuation

DRS: 14 VMs off esx-dc1-14 · CPU ready 14%

Traced to the maintenance mode that moved the Kubernetes nodes onto a busy host.

Resolve

Slow writes on one datastore

ds-prod-ssd-03 write latency 48 ms · snapshot chain of 6

A snapshot chain on the database VM, not the storage array.

Resolve

A memory-starved VM

mem.vmmemctl 6.1 GB · swapped 2.4 GB · pool limit

A resource pool limit, set last week, is ballooning the VM.

Prevent

A datastore filling up

93% used · +6 GB/min · full at 10:12

Flagged 48 minutes before VMs on it pause for lack of space.

Prevent

A certificate near expiry

vcsa-prod-01 machine SSL · VMCA · 12 days left

Renewal staged for the vSphere team before it expires.

Prevent

A host hardware fault building

esx-dc1-09 · DIMM A4 · correctable errors climbing

Flagged so the host is evacuated in a quiet window, not at peak.

Operate

Snapshots nobody removed

37 snapshots older than 7 days · 2.8 TB

Listed by VM and owner for the next cleanup window.

Operate

Alarms that never clear

Datastore usage on disk · Red on 11 datastores

Thresholds tuned per datastore, so a red alarm means act now.

Operate

Hosts out of step

cl-prod-02 · 2 of 8 hosts on an older ESXi build

Listed for the next patch window, cluster by cluster.

Access

What access does NeuBird need in vCenter?A view of the inventory, and nothing more.

NeuBird signs in to vCenter with one account that can view your inventory. It can read everything that permission covers and change nothing. When the fix belongs in vSphere, NeuBird stages it for your team, and actions in other connected systems follow the Suggest, Recommend and Act policy you set per action, with configurable human approval gates.

NeuBird reads

  • Inventory, alarms, events and tasks
  • Performance metrics and host logs
  • Snapshots and certificates
  • Only the objects your vCenter permission covers

Through vCenter, NeuBird never

  • Power a VM or host on or off
  • Change host, cluster or VM configuration
  • Create, revert or delete snapshots
  • Start a vMotion or put a host in maintenance mode

Set up in three steps

  1. 1Create a vCenter account, such as neubird@vsphere.local.
  2. 2On the vCenter Server root object, give it a view-only role and tick Propagate to children.
  3. 3In NeuBird, add the connection with the vCenter URL, username and password.

One connection per vCenter Server, over HTTPS. NeuBird is deployable in under one hour.

FAQ

Common questions

What does NeuBird do for VMware vCenter?

NeuBird is the AI SRE. It reads your vCenter inventory, alarms, events, performance metrics, host logs, snapshots and certificates alongside your application telemetry, Kubernetes and change history, and traces an incident from the application down to the host or datastore underneath it. It isolates root cause in under 5 minutes at 94% accuracy, flags datastore, snapshot, certificate and host hardware risks 30 to 60 minutes before impact, and runs the hygiene between incidents.

What vCenter data can NeuBird read?

The vSphere inventory (datacenters, clusters, hosts, virtual machines, datastores and resource pools), active alarms, events and tasks such as DRS migrations and maintenance mode, performance metrics such as CPU ready, memory ballooning and swapping, and datastore latency, host logs such as vmkernel.log and hostd.log, snapshots with their age and size, and certificates with their expiry dates.

Can NeuBird change anything in vCenter?

No. Through the vCenter connection NeuBird cannot power on, change or delete anything: no power operations, no configuration changes and no snapshot changes. When the fix belongs in vSphere, NeuBird stages it as a recommendation for your team. Actions in other connected systems follow the Suggest, Recommend and Act policy you set per action, with configurable human approval gates.

What access does the vCenter account need?

Create a vCenter Single Sign-On or Active Directory account, such as neubird@vsphere.local, and on the vCenter Server root object assign it a role that grants view access only, with Propagate to children ticked. Objects outside that permission are invisible to NeuBird. If host log queries return a permission error, clone that role, add the Global > Diagnostics privilege, and assign the clone instead.

How does NeuBird reach a vCenter Server on a private network?

NeuBird connects to vCenter over HTTPS on port 443. If vCenter presents a certificate from an internal certificate authority, including the VMware Certificate Authority (VMCA), your NeuBird operator adds that CA to the deployment. A vCenter Server on a private network can be reached through a NeuBird component deployed in that network, and the connection keeps its own endpoint and credentials.

Does NeuBird replace VMware Aria Operations?

No. VMware Aria Operations is a separate NeuBird connection that brings its alerts, symptoms, recommendations, metric statistics and capacity into investigations. Connect both, and NeuBird correlates Aria Operations analytics with live vCenter inventory, events and host logs in the same investigation.

Does NeuBird store our vCenter data?

No. NeuBird queries vCenter in place, with no bulk telemetry stored. Its memory keeps conclusions, such as a verified root cause, the evidence it cited and the fix your team approved, never raw metrics, events or logs.

How many vCenter Servers can NeuBird connect to?

One connection covers one vCenter Server, so add one connection for each vCenter you run. Each connection needs only the vCenter URL, a username and a password, and a healthy connection reaches Synced in NeuBird. NeuBird is deployable in under one hour.

NeuBird for VMware vCenter

From the slow service to the host underneath it.

See NeuBird, the AI SRE, trace an incident from the application through Kubernetes to vCenter, with the fix staged for your team.