ADR 0015 — Help Center / Learning Hub (Meraki-inspired)¶
- Status: Accepted (locked 2026-05-08, formalized 2026-05-09)
- Date: 2026-05-08
- Deciders: TLSStress.Art project
- Targets: v4.7+ (skeleton in next sprint; content authoring ongoing across releases)
- Source memo:
discuss_help_center_2026_05_08.md
Context¶
Operators landing in the dashboard for the first time face a
high-cognitive-load surface — 9 NGFW vendors × 3 catalog kinds × 5
inspection profiles × 3 topology modes × N test plans. Without
in-app guidance the bench leaks into "tribal knowledge held by
whoever installed it last", which violates the
project_quality_excellence_policy_2026_05_08 standard.
User confirmed 2026-05-08 with a Cisco Meraki sidebar screenshot as the reference design. Meraki's Help Center is one of the few network-product UIs where on-page help, glossary, primers, how-tos, and AI Q&A all live in the same right-rail surface — operators never leave the workflow to learn the workflow.
Decision¶
Ship a Help Center as a first-class right-rail surface in the dashboard, modeled on the Meraki sidebar. Composed of 7 components, all queryable from the existing Find palette (Cmd+K) for cross-cutting search:
- Cmd+K palette — already shipped (Find Palette PR-1..PR-5). Glossary entries surface here in PR-6 of Find Palette.
- Tooltips — short context bubbles on every non-obvious UI element, linkable to a Glossary entry.
- Glossary — 25+ initial terms (TLS Inspection, BGP
saturation, Quadrant decomposition, RFC 9411, NetSecOPEN,
Inspection Profile, etc.). YAML-source-of-truth in
platform/help-center/glossary/. - Primers — 8+ initial primers. One-page concept walkthroughs (e.g. "What's a quadrant decomposition?", "Why does BGP saturation matter?"). Markdown-source-of-truth.
- How-tos — 10+ initial workflows. Step-by-step playbooks (e.g. "Configure FTD via FMC for TLS decryption", "Run a capacity-fitted BGP saturation test against a Cisco FTD CSF6160"). Markdown-source-of-truth.
- Videos — short (1-3 min) walkthroughs hosted via Mux (consistent with marketing site infra).
- AI Q&A — natural-language Q&A grounded in the above content via vector embeddings + retrieval-augmented generation. LLM choice deferred (see Open Questions in source memo).
Sidebar layout (Meraki-mapped): Help Center / Community / Learning Hub / DUT info / API docs / Marketplace / What's New.
Roadmap¶
| PR | Scope | Lands |
|---|---|---|
| Help Center PR-1 | Skeleton route + glossary YAML schema + 3 langs (pt-BR / en / es) | next sprint |
| Help Center PR-2 | Tooltips component + 5 highest-impact insertions | next sprint |
| Help Center PR-3 | Glossary content (25 terms) + Find Palette PR-6 wiring | +1 sprint |
| Help Center PR-4 | Primers (8 initial) | +2 sprints |
| Help Center PR-5 | How-tos (10 initial) + Videos shells | +2 sprints |
| Help Center PR-6 | AI Q&A endpoint + RAG pipeline | +3 sprints (depends on LLM decision) |
Open questions¶
Tracked in source memo discuss_help_center_2026_05_08.md:
- Community channel — Discord / GitHub Discussions / Slack /
Discourse / hybrid? (cross-ref
project_community_strategy) - Marketplace timing — alongside Help Center or after Certification Program ships?
- LLM choice for AI Q&A — Claude / GPT-4 / open-source local?
- Translation pipeline — manual + glossary-locked vs LLM-assisted?
Consequences¶
- Positive: lower onboarding friction; reduces support-ticket volume; differentiates against IXIA's documentation-as-PDF approach.
- Negative: ongoing content authoring debt; LLM cost (RAG inference) when AI Q&A ships.
- Reversible if needed: each component is independent; we can ship 1-3 and delay the rest if scope pressure demands.
References¶
- Source memo:
discuss_help_center_2026_05_08.md - Cross-ref:
project_quality_excellence_policy_2026_05_08.md - Cross-ref:
project_community_strategy_2026_05_08.md - Cross-ref:
project_dashboard_find_palette_2026_05_08.md