10.3 X-Ray, Collector, and the Objective 1.3 Tool Inventory
Key Takeaways
- X-Ray is an automated testing framework that benchmarks HCI performance and resilience, running scenarios such as node failure during load rather than raw synthetic I/O alone.
- Collector gathers inventory and utilization data from an existing VMware environment to support Nutanix sizing before any migration begins.
- Collector runs pre-sales and pre-migration; Move runs the migration itself, so the two are sequential rather than alternatives.
- Objective 1.3 names exactly seven tools: Move, NGT, Foundation and Foundation Central, the Witness Service/VM, Self-Service Restore, X-Ray, and Collector.
- Each of the seven maps to one distinct phase: sizing, deployment, migration, guest integration, arbitration, self-service recovery, and validation.
10.3 X-Ray, Collector, and the Objective 1.3 Tool Inventory
Objective 1.3 names seven tools. Five of them appear elsewhere in this guide where their function naturally belongs. This section teaches the remaining two — X-Ray and Collector — and then assembles all seven into one recognition table, because the most common way this objective is tested is a straight "which tool would you use" scenario.
X-Ray: Benchmarking Performance and Resilience
X-Ray is an automated testing framework for hyperconverged infrastructure. It is free, deploys as a virtual appliance, and runs test scenarios against a target cluster.
What makes X-Ray distinctive is the second half of its remit. Conventional storage benchmarking tools measure throughput and latency under a synthetic load and report a number. X-Ray does that, but it also tests resilience — how the platform behaves when something breaks while the load is running.
Representative scenarios:
- Performance under a defined workload profile, such as an OLTP-like or VDI-like mix, reported as throughput and latency over time.
- Node failure during load. X-Ray fails a node mid-test and charts what happens to application performance while the cluster rebuilds. A platform that posts an excellent peak number but collapses during a rebuild looks very different under this test.
- Rolling upgrade during load, showing the performance impact of a non-disruptive upgrade on running workloads.
- Noisy neighbour, showing whether one aggressive VM degrades others.
The results are charted over the test duration rather than reduced to a single figure, which is the point: the shape of the curve during a failure is the information a synthetic benchmark hides.
[!TIP] Use-case sentence to remember: X-Ray is how you compare HCI platforms — or validate your own cluster — on both performance and behaviour during failure. If a scenario mentions evaluating vendors, proving resilience claims, or measuring the impact of a node failure on running applications, the answer is X-Ray.
Collector: Sizing From the Existing Estate
Collector answers the question that comes before any Nutanix purchase: what do we actually have today, and how big does the Nutanix cluster need to be?
Collector connects to an existing VMware environment — typically vCenter — and gathers inventory and utilization data: how many VMs there are, their configured CPU, memory, and disk, and crucially their actual consumption over a collection period.
That distinction between configured and consumed is the whole value. Estates are routinely over-provisioned, and sizing a new platform from configured values produces a cluster far larger and more expensive than necessary. Collector supplies measured data so sizing reflects real demand.
Characteristics to recognize:
- It is a free, read-only data-gathering tool. It changes nothing in the source environment.
- It runs before any migration — during evaluation and design.
- Its output feeds Nutanix sizing for the proposed cluster.
Collector and Move Are Sequential, Not Alternatives
A frequent point of confusion, and a good distractor:
| Phase | Tool | Question it answers |
|---|---|---|
| Evaluate and size | Collector | What do we have, and how much Nutanix do we need? |
| Deploy | Foundation | How do these nodes become a cluster? |
| Validate | X-Ray | Does it perform and survive failure as claimed? |
| Migrate | Move | How do the workloads get onto AHV? |
Collector never moves a VM. Move never sizes a cluster.
The Complete Objective 1.3 Tool Inventory
All seven tools, each mapped to the phase it serves and to where this guide teaches it:
| Tool | Purpose | Phase | Covered in |
|---|---|---|---|
| Collector | Gathers inventory and real utilization from an existing VMware estate to size Nutanix | Pre-sales / design | This section |
| Foundation / Foundation Central | Images bare-metal nodes and creates clusters; Foundation Central does it remotely at scale | Deployment | Section 10.2 |
| X-Ray | Automated performance and resilience testing of HCI | Validation | This section |
| Move | Migrates VMs onto AHV from ESXi, Hyper-V, AWS, and Azure | Migration | Section 10.1 |
| NGT (Nutanix Guest Tools) | In-guest agent enabling application-consistent snapshots, VM mobility, and self-service restore | Guest integration | Section 6.3 |
| Self-Service Restore (SSR) | Lets a VM owner recover their own files from a snapshot without an administrator | Day-2 recovery | Section 6.3 |
| Witness Service / VM | Arbitrates leadership for two-node clusters and Metro deployments to prevent split brain | Availability | Sections 2.1 and 9.2 |
[!IMPORTANT] Two of these depend on each other and are frequently paired in questions. Self-Service Restore requires NGT to be installed and enabled in the guest, because SSR works through the in-guest agent. A question that describes a user unable to restore their own files from a snapshot is very often an NGT question in disguise.
Quick Discrimination Drill
Work through these until the mapping is automatic:
- Prove that application latency stays acceptable when a node fails mid-workload. → X-Ray
- Determine how much CPU and memory the existing 900 VMs actually consume. → Collector
- Turn 12 new bare-metal nodes into a cluster. → Foundation
- Do the same at 40 stores without travelling. → Foundation Central
- Move those 900 VMs from vSphere onto AHV. → Move
- Take an application-consistent snapshot of a SQL Server VM. → NGT
- Let a user recover a file they deleted this morning, themselves. → SSR, which requires NGT
- Stop a two-node branch cluster from splitting brain. → Witness VM
An organization is comparing two hyperconverged platforms and wants evidence of how application performance behaves while a node fails and the cluster rebuilds. Which Nutanix tool produces that evidence?
Before proposing a Nutanix cluster, a partner needs to know what the customer 900 existing VMware virtual machines actually consume rather than what they are configured for. Which tool is used?
A VM owner reports that they cannot recover their own deleted files from a snapshot using Self-Service Restore. What is the most likely prerequisite that has not been met?