chadstthomas commited on
Commit
2ae69c2
·
verified ·
1 Parent(s): 87ca5df

Initial release: nutanix-knowledge (505 Q&A pairs)

Browse files
Files changed (2) hide show
  1. README.md +116 -0
  2. data/train.jsonl +0 -0
README.md ADDED
@@ -0,0 +1,116 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: cc-by-4.0
3
+ task_categories:
4
+ - question-answering
5
+ language:
6
+ - en
7
+ size_categories:
8
+ - n<1K
9
+ pretty_name: Nutanix Product, Licensing and Support Knowledge (Q&A)
10
+ tags:
11
+ - nutanix
12
+ - aos
13
+ - ahv
14
+ - prism
15
+ - nc2
16
+ - nci
17
+ - licensing
18
+ - technical-support
19
+ - knowledge-base
20
+ configs:
21
+ - config_name: default
22
+ data_files: data/train.jsonl
23
+ ---
24
+
25
+ # Nutanix Product, Licensing & Support Knowledge (Q&A)
26
+
27
+ A curated question-and-answer dataset covering the **Nutanix platform** (AOS, AHV, Prism,
28
+ NC2, Files/Objects/Volumes, NDB, DR, LCM and the wider Nutanix Cloud Platform portfolio),
29
+ **Nutanix licensing** (NCI editions, term/per-core models, NC2 consumption), and **opening
30
+ and managing support tickets with Nutanix**. 505 entries, each grounded in an official
31
+ Nutanix source URL.
32
+
33
+ ## Motivation
34
+
35
+ Support and operations teams working with Nutanix environments need accurate,
36
+ source-attributed answers about current product naming (Nutanix Disaster Recovery, Nutanix
37
+ AI Factory / Enterprise AI, NUS/NDB), edition-based licensing, and the support-case
38
+ workflow. Public training corpora lag the current portfolio and support-policy details.
39
+
40
+ ## Dataset structure
41
+
42
+ One JSON object per line in `data/train.jsonl`:
43
+
44
+ | Field | Type | Description |
45
+ | --- | --- | --- |
46
+ | `id` | string | Stable id (`ntx-0001` … `ntx-0505`) |
47
+ | `question` | string | The question |
48
+ | `answer` | string | Grounded answer (plain text, 40–130 words typical) |
49
+ | `product` | string | Product/area tag (AOS, AHV, Prism Central, NC2, NDB, Files, Nutanix Support, …) |
50
+ | `category` | string | `architecture` / `product-catalog` / `licensing` / `support-ticketing` / `operations` |
51
+ | `difficulty` | string | `basic` / `intermediate` / `advanced` |
52
+ | `source` | string | Official URL the answer was grounded in |
53
+
54
+ ### Category distribution
55
+
56
+ | Category | Entries |
57
+ | --- | ---: |
58
+ | product-catalog | 108 |
59
+ | architecture | 105 |
60
+ | operations | 105 |
61
+ | support-ticketing | 94 |
62
+ | licensing | 93 |
63
+
64
+ Difficulty mix: 160 basic / 205 intermediate / 140 advanced.
65
+
66
+ ## Coverage highlights
67
+
68
+ - **Architecture** — AOS/Distributed Storage Fabric, data locality and tiering, RF2/RF3
69
+ and EC-X, metadata-based snapshots, AHV, hypervisor choice, Prism Element vs Prism
70
+ Central (single-VM and 3-VM deployments), replication modes (async / NearSync 1–15 min /
71
+ sync RPO 0), LCM inventory vs update, Foundation Central, NC2 clouds (AWS, Azure, Google
72
+ Cloud, OVHcloud), NX and OEM hardware (Dell XC, Lenovo HX, Cisco, HPE, Fujitsu).
73
+ - **Products** — Nutanix Cloud Platform portfolio naming (NCI, NUS, NDB, Nutanix
74
+ Enterprise AI / AI Factory, NC2, NCM, Nutanix Central), Files/Objects/Volumes details,
75
+ Flow Network Security, NDB engine list, Prism Pro/Ultimate (AIOps tiers), Move.
76
+ - **Licensing** — Starter/Pro/Ultimate editions and feature gating, capacity-based
77
+ per-core/SSD licensing with 1–7 year terms, VDI per-user and ROBO per-VM models,
78
+ appliance licensing (non-transferable), AHV/Prism inclusion, Files 1 TiB / Objects 2 TiB
79
+ free allowances, NC2 PAYG (published $/core-hour rates) vs BYOL portability.
80
+ - **Support & ticketing** — portal Open Case flow (serial number, hypervisor + AOS
81
+ version), P1–P4 definitions and targeted initial response levels (P1: 1 hour; 30 minutes
82
+ under Mission Critical Support), escalation via SRE / Head of Support, Pulse/Alerts
83
+ proactive monitoring, EOSL policy, hardware maintenance and RMA/dispatch terms, RFE
84
+ target.
85
+ - **Operations** — LCM dark-site bundles, PC/PE component support, cluster grow/shrink,
86
+ DR protection/recovery plans and test failovers, NDB lifecycle, Files/Objects day-2
87
+ concepts, support-case ops.
88
+
89
+ ## Method and sourcing
90
+
91
+ Entries were authored against official Nutanix public sources — `nutanix.com` product and
92
+ support pages (including the Nutanix Support Policies and Support Program Guide), the
93
+ official edition-comparison pages, and NC2 pricing pages — with each entry citing its
94
+ source. Details that live behind the authenticated support portal (exact CLI/NCC command
95
+ syntax, Prism menu paths, upgrade click sequences, version matrices) are **not asserted**:
96
+ where relevant, entries state what is not publicly documented and point to the portal
97
+ documentation. Six CLI-adjacent entries include explicit version-sensitivity caveats.
98
+
99
+ ## Caveats
100
+
101
+ - **Time-sensitive values** (PAYG $/core-hour rates, policy targets, edition feature
102
+ lists) reflect official pages as reviewed on **2026-10-30**. Re-verify before
103
+ quoting or contracting.
104
+ - Product lifecycle items that Nutanix has not formally documented publicly (e.g., formal
105
+ "Leap" rename statements, Frame/Mine commercial status, Hyper-V retirement dates) are
106
+ answered with explicit uncertainty rather than guessed.
107
+ - The support-policy response targets are policy objectives, not resolution guarantees.
108
+
109
+ ## Intended use
110
+
111
+ - Fine-tuning / SFT for infrastructure-support assistants.
112
+ - Retrieval-augmented generation (RAG) knowledge base for Nutanix operations.
113
+ - Evaluation of LLM knowledge on the current Nutanix portfolio.
114
+
115
+ Not intended for: contractual licensing decisions, SLA commitments, or as a replacement
116
+ for official Nutanix documentation.
data/train.jsonl ADDED
The diff for this file is too large to render. See raw diff