release: promote Slice 0 foundation - #51
Merged
Merged
Conversation
…ion doc Reshape the Sith plan to the research-backed local-first dual-mode direction and reconcile the two research lanes into one authoritative synthesis. - research/USE-CASE-AND-SHAPE.md: reconcile the evidence set (PR #16) with the sharper day-0 framing (PR #15) — local-first single-binary CLI/TUI as the day-0 wedge, MCP server in v1, honest wedge-sizing — into one authoritative document. - CHARTER: two wedges (local client = adoption, governed federation = moat); individual operator as top-of-funnel; an adoption success criterion. - SCOPE: scope OUT "another single-cluster console", scope IN "aggregated multi-cluster local client" (fixes the Lens contradiction); contractual non-goals. - ARCHITECTURE: three run-modes on one Go binary; source-abstract read federation (local kubeconfig OR OCM spoke); four-mode identity; connector framework; cost overlay; cosign/SLSA/SBOM; SPIFFE-without-SPIRE. - ROADMAP: Phase L (day-0 local mode, no OCM) ahead of the hub track; MCP read in Phase L; MCP write behind the governed write path. - EPICS: add E11 (local client), E12 (connector framework), E13 (cost overlay), each with per-feature flowcharts; E2 made source-abstract; E7/E8/E9 reshape notes; updated roadmap map + open questions. - SITH-NOTION.md: one self-contained Notion-ready page — overview, wedge, form factor, prioritized roadmap, carry/discard/net-new, every epic with a flowchart per feature, and the open questions.
Add the architect-owned build plan for the local-first Phase-L wedge (the single `sith` local fleet client) so a fresh builder session can start Slice 0 with no extra context: - docs/BUILD-SEQUENCE.md — the locked, ordered slice plan (0 foundation → F2.1+F11.1 → F11.2 → F11.5 → F11.3 → F11.6 → F7.1, plus a parallel packaging track), each slice mapped to its issue(s), with dependencies, acceptance criteria, and the Q12–Q15 dependency summary. - docs/specs/SLICE-0-foundation.md — a self-contained implementation spec: module init, cobra CLI skeleton, config/slog, the F2.1 fleet-source seam with a stub, Makefile, green GitHub Actions CI, .golangci.yml, and the exact test list and acceptance criteria. - docs/CONVENTIONS.md — branch model, Conventional+DCO+SSH-signed commits with the GSTACK-Checkpoint trailer and no AI attribution, CI merge gates, Go style, and test requirements. - sessions/ — the GSTACK session-journal scaffold (README + template) and this planning session's journal. Grounded in docs/SITH-NOTION.md (E2/E7/E9/E11, roadmap, open questions) and issues #29 #38 #32 #33 #34 #35 #36 #37 #39. Records one deliberate divergence from roadmap #39 (E9 packaging folded to a parallel, non-gating track) and the Q12/Q14/Q15 defaults. Planning only — no product code. GSTACK-Checkpoint: 2026-07-10/build-plan-architecture#1 Signed-off-by: Gnani Rahul <89947795+gnanirahulnutakki@users.noreply.github.com>
…kflow Add a "Who this is for — the target user's daily surface" section mapping GR's day-to-day stack (Kubernetes, Helm, ArgoCD, Docker, Python/bash, Fluentd/Fluent-bit, Grafana/Prometheus, multi-cloud AWS/Azure/GCP, vuln fixes, cloud networking) onto the plan, and a concrete "User-workflow fit" line to each slice tying it to a real task (morning fleet sweep, incident triage + vuln sweep, debug the failing pod, GUI/share, run safely on a corp laptop, ask the agent, install like kubectl). Slice 0 stays workflow-agnostic. Leaves an anchored hook for a fuller GR-workflow profile (to land as docs/GR-WORKFLOW-PROFILE.md) that will refine the per-slice fit lines and the E12/E13 connector priority order once supplied. Planning only — no product code. GSTACK-Checkpoint: 2026-07-10/build-plan-architecture#2 Signed-off-by: Gnani Rahul <89947795+gnanirahulnutakki@users.noreply.github.com>
Design the seam every integration plugs into: the 7-verb connector contract (discover/read/query/diff/plan/execute/verify) as idiomatic Go, translated from the incident-derived corpus dataclasses. - Capability-scoped interfaces (Reader/Differ/Planner/Executor/Verifier) so read-only sources implement only what they can do and the compiler proves what a connector supports. - Capability enum + Registry with fail-safe registration (a declared capability must implement its interface) and capability-based lookup; one canonical connector per tool, closed three-kind taxonomy (E12). - Normalized source-abstract model: ResourceRef, Evidence/Provenance, Fact, Query/Selector, Coverage, and the operational-graph shape (typed relations) that makes cross-cluster correlation first-class. - First adapter: local-kubeconfig as a Reader-only read adapter (discover/read/query) — list contexts, fan out across contexts, surface unreachable, credentials never leave the machine. - AsSource bridge keeps Slice-0's fleet.Source seam intact (no redefinition); Slice 1 swaps one injection line. - W1-W4 additivity table: adding a connector is one package + one Register call, zero changes above the seam. Spec-only; a builder implements Slice 1 (#38 + #32). Open decisions recorded in the doc, not blocking. Refs: #38, #20 Signed-off-by: Gnani Rahul <89947795+gnanirahulnutakki@users.noreply.github.com>
…le stakes (F11.5) Design/spec for the day-0 adoption surface: the `sith` CLI + TUI cache-first fleet render (F11.2 #33) and per-pod table stakes (F11.5 #35). Sits above the Slice-0 foundation and consumes Slice-1's local-kubeconfig adapter. Covers: - CLI + TUI information architecture (k9s-informed, fleet-native — cluster is a filter dimension, not a mode; fleet-first default scope). - Cache-first render model: store as single render source, <100ms interactions, <250ms first paint, cold/warming/warm/degraded/offline states, per-cell and per-cluster staleness, honest coverage line on every aggregate. - Resource lenses ordered to daily K8s ops (pods/deployments/events/nodes first). - Per-pod actions (logs tail+follow, exec, port-forward, describe, get-yaml) each mapped onto Architect A's 7-verb source-adapter contract (read triad discover/read/query + streaming logs/exec/port-forward); local-identity, never-governed boundary asserted. - Cross-cluster search + correlation grammar and navigation. - Screens/states catalog, full keybindings, testable acceptance criteria per Slice 2 / Slice 3, and recorded open decisions (D1 YAML-edit write capability is the one item flagged back to Architect A). Signed-off-by: Gnani Rahul <89947795+gnanirahulnutakki@users.noreply.github.com>
…integration waves Add the design/spec that sits above the Slice-0 foundation and beside the 7-verb connector seam: it grows E2's normalized fleet model into a four-lens operational graph (live/desired/timeline/telemetry) correlated by OTel-style EntityRef keys, specs the rule-based Investigation Brain (symptom -> scope -> evidence -> hypotheses -> validation -> plan) with a six-rule hypothesis catalog for the failures GR actually hits, and scores every integration across four waves by verb subset, lenses fed, and effort tier. Keeps the retention guardrail (no telemetry lake): timeline is a bounded change ring, telemetry is query-through against the backend's own retention. Draws the local-wedge vs hub/federation line: one brain, two modes -- advisory command in local, governed typed-intent proposal through the PEP (E4/E5) in hub. Proposes a new epic E14 for the brain and E2 features F2.6/F2.7 for the graph substrate as an open decision. Includes acceptance criteria for all three parts. Signed-off-by: Gnani Rahul <89947795+gnanirahulnutakki@users.noreply.github.com>
Current (mid-2026) competitive and standards scan feeding epic creation. Builds on the prior research and the just-landed specs (PRs #40-#43); does not duplicate them. - 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), and IDPs. - Category framing: local-first, cross-cluster-by-default, no-account is a real (beachhead) wedge; the deterministic rule-based Investigation Brain is the differentiator vs LLM AI-SRE (the hallucination gap + third-party data egress). - Standards to align with: OpenTelemetry k8s semconv (RC), Kubernetes API conventions, OCM, MCP, SLSA / cosign / SBOM, client-go exec-plugin auth. - Differentiation and positioning + an explicit "what NOT to build" list. - Recommended epic themes mapped to existing epics #18-#39, with one new epic to open: E14 Investigation Brain. Every external claim carries a URL; search-surfaced items are flagged. Signed-off-by: Gnani Rahul <89947795+gnanirahulnutakki@users.noreply.github.com>
Fold the July-2026 market-research findings into the roadmap doc so the epic backlog and the roadmap tell one story: - E14 Investigation Brain sequenced across phases (advisory subset in Phase L, deterministic four-lens brain in P1, governed-plan in P2/P3). - Phase-L build sequence table (Slices 0-6 + P) mapped to issues. - Integration waves W1-W4 (E12) with the daily-core coverage the six brain rules need. - Standards-alignment gates (MCP 2026-07 auth + enforce-at-execution, OTel semconv keys, client-go ExecCredential v1, SLSA/cosign/SBOM, K8s API conventions). - Anti-scope: no LLM in the root-cause critical path; no act-from-chat. Signed-off-by: Gnani Rahul <89947795+gnanirahulnutakki@users.noreply.github.com>
Pin the supported Go baseline, minimal runtime dependencies, and strict formatting, linting, and security checks for Slice 0. GSTACK-Checkpoint: 2026-07-10/slice-0-foundation#1 Signed-off-by: Gnani Rahul <gnani.nutakki@gmail.com>
Implement build metadata, fail-safe config precedence, structured logging, and the source-abstract FleetResult stub seam with table-driven tests. GSTACK-Checkpoint: 2026-07-10/slice-0-foundation#2 Signed-off-by: Gnani Rahul <gnani.nutakki@gmail.com>
Wire Cobra commands to the stub fleet source, provide deterministic text and JSON output, and verify the real binary boundary with subprocess smoke tests. GSTACK-Checkpoint: 2026-07-10/slice-0-foundation#3 Signed-off-by: Gnani Rahul <gnani.nutakki@gmail.com>
Run module verification, formatting, vet, pinned lint, build, race tests, coverage, and process-level e2e checks with least-privilege action permissions. GSTACK-Checkpoint: 2026-07-10/slice-0-foundation#4 Signed-off-by: Gnani Rahul <gnani.nutakki@gmail.com>
Replace the planning-only README with runnable build guidance and capture implementation, validation, review, and commit checkpoints in the GSTACK journal. GSTACK-Checkpoint: 2026-07-10/slice-0-foundation#5 Signed-off-by: Gnani Rahul <gnani.nutakki@gmail.com>
The v9 action invokes golangci-lint run itself, so provide only the ./... package pattern and record the failed remote run in the GSTACK journal. GSTACK-Checkpoint: 2026-07-10/slice-0-foundation#6 Signed-off-by: Gnani Rahul <gnani.nutakki@gmail.com>
Exercise the real make build path with deterministic ldflags and assert the compiled binary reports the injected version, commit, and date. GSTACK-Checkpoint: 2026-07-10/slice-0-foundation#7 Signed-off-by: Gnani Rahul <gnani.nutakki@gmail.com>
feat(cli): land Slice 0 foundation walking skeleton
docs(build): lock Phase-L build sequence, Slice-0 spec, and conventions
docs(e2): F2.1 source-abstract fleet model + connector/source-adapter contract (#38)
docs(spec): local fleet UX — cache-first render (F11.2) + per-pod table stakes (F11.5)
docs(e2): read-federation four-lens graph, investigation brain, integration waves
docs(research): 2026 market & landscape research for Sith
docs(roadmap): consolidate July-2026 research into the combined roadmap
This was referenced Jul 10, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Release increment
Promote the first coherent tested Sith increment from
devtomain:sithbinary withversion, typed-emptyclusters,ui, andhubfleet.Sourceseam, GSTACK journals, and strict CIEvidence
devas3a20e2ddevtip:a9bf340This is the release checkpoint required after Slice 0. No tag or public package is cut yet; full cosign/SLSA/SBOM packaging remains Slice P / E9 #27.