docs(research): 2026 market & landscape research for Sith

[gnanirahulnutakki]

What this is

A current (mid-2026) market & landscape research pass for Sith, written to feed epic creation. One new file: docs/research/MARKET-RESEARCH-2026-07.md (~3,960 words, cited). Builds on the prior research and the just-landed specs (PRs #40–#43); does not duplicate them.

Contents

• Competitive landscape + comparison matrix across four bands — local clients (k9s, Lens/FreeLens, Headlamp, kubevious, Aptakube), platforms (Rancher, Portainer, Devtron, ArgoCD UI), AI-SRE/auto-triage (k8sgpt, HolmesGPT, Robusta, Botkube, Cleric, Komodor, Cast AI), IDPs. Each with what-it-does, gaps, and real-user signal (k9s multi-cluster issues #1006/#2730/#3374/#1430; ArgoCD "you lose the single dashboard"; Portainer's missing drift detection; k8sgpt's OpenAI egress).
• Category framing — is local-first, cross-cluster-by-default, no-account a real wedge? Verdict: yes as a beachhead; the rule-based Investigation Brain is the differentiator.
• Standards to align with — OTel k8s semconv (flagged: release-candidate, not yet stable), K8s API conventions, OCM, MCP, SLSA/cosign/SBOM, client-go exec-plugin auth.
• Differentiation & positioning — default-tool case, strongest wedge + sequence, risks/objections, and an explicit what NOT to build list.
• Recommended epic themes → existing epics E0: OCM substrate and falsification #18–Sith implementation roadmap #39, plus one new epic to open.
• Cited bibliography with a verification note (search-surfaced items flagged).

Headline findings

• Top competitors that matter: k9s (the on-ramp incumbent, won't do multi-cluster), Headlamp (CNCF dual-mode client), Aptakube (only aggregated multi-cluster — closed/paid), and the AI-SRE band (HolmesGPT/k8sgpt/Cleric) — all LLM-based, mostly SaaS/single-cluster.
• Wedge verdict: real. The empty row in the matrix is local-first + cross-cluster + four-lens + rule-based RCA + governed action, in one OSS binary. No incumbent occupies it.
• The differentiator vs the AI-SRE band is the hallucination gap: LLMs root-cause without live evidence and sound confident when wrong. Sith's brain reasons over the live four-lens graph, cites signals, abstains on missing coverage, and never sends cluster data to a third-party model.
• New epic recommended: E14 — Investigation Brain (already proposed in the E2 spec §3.7). Highest-leverage addition; local advisory subset in Phase L, governed-plan mode with the write path.

Notes

• Docs-only; no product code.
• Commit is SSH-signed + DCO signed-off, ArdurAI signing identity, no AI attribution.
• Do not merge — for review.