View on GitHub

OdbDesign ODB++ Parser

REST API and C++ library for parsing ODB++ Design archives

Program status — server + 3D client + info client (2026-11-02)

   
Purpose Full status report across the three main projects: recently completed, in flight, what remains, and one dependency-ordered action plan interleaving all three. Successor to program-status-2026-11-01.md; covers the delta since then plus the standing backlog. Itemization level only — execution plans get cut from this, not into it.
Verified 2026-11-02 against origin, all three repos fetched (--prune) first; PR/issue claims re-checked via gh; cluster probed live and read-only via kubectl (context k3s-debian13vm). Line-level cites re-read locally this session: m6-4-client-fanout.md, README.md, info-client DesignService.cs / service.proto / IOdbDesignRestApi.cs / ConnectionService.cs, 3D new-issues-remaining.md / mirrored_text/ANALYSIS.md / packaging-and-release-plan.md / opt/client-grpc-optimization-remaining-work.md §4. Also run this session: a REST-first claim sweep across all three repos, git check-ignore on the info client’s obj/ codegen, git stash list (3D), and git status --porcelain re-checks. No builds or tests were run — read-only scans.
Repo heads at verification Server: origin/development @ 59ed1cd (local checkout = tip, clean at scan time — this document itself is now that tree’s sole untracked file; see “own landing” before §4) · origin/nam20485 @ 99f03f3 (91 ahead / 1 behind) · staging @ 9748709 · main @ 6074592 · release @ 33448b5 (down-chain frozen since mid-September). 3D: origin/development @ e455dd7 (PR #67 merge) · origin/main @ 0430d86 (3 behind, 0 ahead). Info: origin/main @ c98a179 — fully promoted.
Cluster (live probe) k3s debian13vm reachable; 18/18 deployments READY 1/1 across 6 namespaces. odbdesign-server-v1 pod on imageID sha256:29a4501d… — the digest the predecessor verified as the #609 build (nam20485-1287), pod start 2026-11-02T00:21:22Z, 0 restarts. SwaggerUI pod up since 2026-08-20 with 60 restarts (last 3d18h ago — investigate-or-waive decision owed, §4 step 4). Argo CD control plane healthy (6 deployments, 56d old) — zero Application CRs cluster-wide; GitOps execution not started.

TL;DR

At a glance

Component Recently completed In flight Headline remaining
Server (nam20485/OdbDesign) #609 merged + tip CI 4/4 green; #609 deployed + live-accepted; docs snapshot committed to development (59ed1cd) M6.6 queued (branch not cut); 7 dependabot + stale #423 open; branch residue + plan-index lag 91-commit promotion nam20485→development (two-way now) → resume frozen chain; M6.6; hygiene (#599, 7× dependabot, SwaggerUI watch); Argo CD execution
3D client (intel-agency/odbdesign-3d-client-prototype) PR #67 MERGED (Nov 2 13:38Z) — stream adoption, review-complete development→main promotion pending (PR not yet open); local checkout unsynced promote to main · NI24 headless stall · #64 Turbot silkscreen · R-2/R-5/R-6/R-7 perf re-measure · packaging decisions
Info client (nam20485/Odbdesign-info-client-india79-b) #27 + #28 + #29: M6.4 protos vendored and promoted to main (Nov 2 00:42Z) — (verified: nothing in flight) GetDesignStream adoption decision (big designs render silently empty today) · stale epics #2–#8 + #10 triage · proto-drift guard

Ecosystem map

flowchart LR
    subgraph SVR["OdbDesign server - nam20485/OdbDesign"]
        direction TB
        NAM["nam20485 @ 99f03f3\n#609 tip - CI 4/4 green Nov 1"]
        DEV["development @ 59ed1cd\n1 ahead / 91 behind"]
        NAM -->|"promotion pending\nnow a 2-way merge"| DEV
    end
    subgraph IMG["GHCR"]
        TAG["odbdesign:nam20485-latest\nfloating tag"]
    end
    subgraph K3S["k3s @ debian13vm - live probe Nov 2"]
        direction TB
        POD["odbdesign-server-v1 1/1\nimageID sha256:29a4501d\n= #609 build - pod start Nov 2 00:21:22Z\n0 restarts"]
        SW["odbdesign-server-swaggerui-v1 1/1\nimageID sha256:4cf342dd\n60 restarts - decision owed step 4"]
        ARGO["Argo CD control plane 1/1\nZERO Applications - GitOps not started"]
    end
    subgraph C3D["3D client - intel-agency"]
        direction TB
        CDEV["development @ e455dd7\nPR #67 merged Nov 2 13:38Z"]
        CMAIN["main @ 0430d86\n3 commits behind - 0 ahead"]
        CDEV -->|"promotion pending"| CMAIN
    end
    subgraph IC["info client - nam20485"]
        IMAIN["main @ c98a179\nfully promoted - 0 open PRs"]
    end
    NAM -->|"docker publish + manual rollout"| TAG
    TAG --> POD
    NAM -->|"proto upkeep push (F6) + M6.4"| CDEV
    NAM -->|"proto upkeep push (F6) + M6.4"| IMAIN
    POD -->|"gRPC GetDesignStream\n+ monolithic GetDesign"| CDEV
    POD -->|"gRPC GetDesign = single content source\nREST = discovery + health only"| IMAIN

Everything still flows from the server’s nam20485 tip: it publishes the container the cluster runs and owns the protobuf contracts both clients vendor. The broken links to fix are the two promotion edges (nam20485→development in the server, development→main in the 3D client) and the missing Argo CD Applications.

Shorthand used above and below: F6 = the clients’ proto-upkeep follow-up — the vendored protoc/ refresh pushed from the server repo (item F6 of the F1–F7 client backlog defined in component-connectivity-client-followups.md); M6.x = milestones of the GetDesign-streaming OOM plan; SI19 = server-issues item 19.


1. Recently completed (delta since 2026-11-01)

The predecessor ended with: “merge #67 → cut M6.6 → commit the two working-tree docs → promote nam20485→development → close #599 + batch dependabot.” One day later: #67 merged, the docs committed (to the wrong line) — everything else is still queued.

flowchart LR
    subgraph S["Server - window Nov 1 to Nov 2"]
        direction TB
        s1["#609 GetDesignStream merged\nNov 1 23:08:08Z - 99f03f3"]
        s2["tip CI 4/4 success\nNov 1 23:08Z"]
        s3["docs snapshot 59ed1cd\ncommitted to development\nNov 2 07:03:49Z"]
        s1 --> s2 --> s3
    end
    subgraph K["Cluster"]
        k1["#609 build rolled out\npod start Nov 2 00:21:22Z\ndigest re-verified live Nov 2"]
    end
    subgraph C["Clients"]
        direction TB
        c1["info #27 protos Nov 2 00:03Z\n#28 + #29 to main Nov 2 00:42Z"]
        c2["3D PR #67 MERGED\nNov 2 13:38:38Z - e455dd7\nfallback + reassembler"]
        c3["connectivity epic closed\nin all 3 repos\nstreaming closed server + 3D\nIC stream adoption still owed"]
        c1 --> c3
        c2 --> c3
    end
    s1 --> k1
    s1 -->|"contract vendored"| c1
    s1 -->|"contract adopted"| c2

1.1 Server — nam20485/OdbDesign

Item Detail
#609 GetDesignStream (M6.1–M6.3) merged — 2026-11-01T23:08:08Z (99f03f3) Chunked GetDesignStream + DesignChunk contract, RESOURCE_EXHAUSTED dispatch guard on over-budget monolithic GetDesign, disk-spool warm cache. Follow-ups rode the trunk: review fixes 5692a4f/a205258/14ef083, F9 heap-trim root cause ae1f159 (warm-RSS gate 5/5), FeatureGroupRecord field-2 fix cc3667a. Window-edge event already recorded by the predecessor; the delta is its aftermath, below.
Tip CI fully green 4/4 workflows success at 2026-11-01T23:08Z on 99f03f3 (CMake-Multi-Platform, Code Coverage, CodeQL, SBOM) — de-risks the 91-commit promotion: the gate is process, not code.
#609 deployed + live-accepted Published as nam20485-1287; the running pod (start 2026-11-02T00:21:22Z, imageID sha256:29a4501d…, 0 restarts) was re-verified live by today’s probe. The live-acceptance result itself (streams OK at 3.31 GiB HWM; over-budget monolithic = clean RESOURCE_EXHAUSTED) is quoted from program-status-2026-11-01.md line 8 — not re-run in this scan.
Program docs committed — to development 59ed1cd (2026-11-02T07:03:49Z) landed program-status-2026-11-01.md (229 lines) + m6-4-client-fanout.md (63 lines, 292 insertions total) straight on development, not the recommended nam20485 line → development is now 1 ahead / 91 behind. This is the mistake the successor must not repeat — see “own landing” before §4.
Dependabot churn Open dependabot PRs dropped 8→7 since 2026-11-01; the closed one was not identified (closed-PR list checked to depth 8 showed only work PRs).

1.2 3D client — intel-agency/odbdesign-3d-client-prototype

Item Detail
PR #67 — M6.4 GetDesignStream adoption MERGED — 2026-11-02T13:38:38Z (e455dd7, base development) ResourceExhausted from monolithic GetDesign → single fallback to GetDesignStream → chunk reassembler (server decides, client never guesses size). Commits 2299e6b + 6dd8592 with all 9 unique three-lens review findings fixed (13 reported → 9 unique; suite 1248/1248 post-fixes per m6-4-client-fanout.md status row). Checks: 6 SUCCESS + trufflehog SKIPPED by design. Upstream feat/getdesign-stream deleted after merge. This closes the streaming epic in the 3D client.
Posture after #67 development = main + 3 commits (2299e6b, 6dd8592, merge e455dd7); development..main = 0 → clean promotion. Zero open PRs; the last 15 workflow runs all success, zero cancellations.

1.3 Info client — nam20485/Odbdesign-info-client-india79-b

Item Detail
M6.4 chain fully promoted — completed 2026-11-02T00:42:05Z #27 (vendored GetDesignStream/DesignChunk contract, protoc/grpc/service.proto:35 and :146-148) merged to nam20485 00:03:31Z → #28 to development 00:12:52Z → #29 to main 00:42:05Z (c98a179). Already known to the predecessor; re-verified in this scan: zero open PRs, clean checkout on nam20485 @ 8679ac8, and origin/main..origin/nam20485 = 0 (the 4-commit reverse drift is promotion merge commits only — nothing stranded).
Contract status Vendored but unconsumed: zero hand-written references. git grep GetDesignStream -- 'src/*.cs' 'tests/*.cs' → 0 — git grep searches tracked files only, and the generated grpc codegen lives under git-ignored obj/ (git check-ignore → .gitignore:32 [Oo]bj/, verified this session). A literal filesystem grep (grep -rn GetDesignStream src tests --include=*.cs) finds exactly 6 lines, all generated codegen in src/OdbDesign.ProductModel/obj/Debug/net10.0/grpc/ServiceGrpc.cs. ConnectionService.cs:67 still calls unary GetDesignAsync (100 MB MaxReceiveMessageSize at :187). Adoption is the open item (§3.3).

1.4 Cross-cutting


2. In progress / actively being worked

flowchart TD
    TIP["S: nam20485 tip 99f03f3\n#609 merged - deployed - accepted"]
    M66["S: M6.6 residency budget\nbranch NOT yet cut\nkResidencyFactor ~5.0\n+ FixtureCapture --stream"]
    PROMOS["S: promotion to development\n91 commits - now 2-way merge\ngate: ruleset 169324 strict checks"]
    HYG["S: hygiene queue\nclose #599 - dependabot #607 #598\nstale #423 - 38-issue triage"]
    RES["S: residue\n3 stale nam branches\nplan-index lag - OOM plan nam20485-only"]
    PROMO3D["3D: development to main promotion\n3 commits - PR not yet opened\ngate: secret-scan flake"]
    SYNC["3D: local checkout unsynced\non merged feat branch - 3 stashes parked"]
    IDLE["IC: nothing in flight\nverified - 0 open PRs - clean tree"]
    TIP --> M66
    TIP --> PROMOS
    PROMOS -.->|"after it lands"| HYG

Nothing is blocked or red — every item waits on a human action (cut, merge, open-PR, triage), not on a fix.

2.1 Server

Item State / gate
M6.6 residency budget — queued, branch not cut Next server PR per m6-4-client-fanout.md §5 and the predecessor §2: re-base cache charges from file size to a resident estimate (kResidencyFactor ≈ 5.0) + FixtureCapture --stream. No nam/m66-residency-budget branch exists on origin. Gate: nothing — cutting from 99f03f3 is the first action. (The spec doc getdesign-streaming-oom-plan.md exists only on nam20485 and was not read in this scan — scope here follows the docs that succeeded it, the m6-4 fan-out plan §5 and the predecessor status §2.)
Open PRs: 8 = 7 dependabot + 1 stale #607 (vcpkg 2026.10.16, updated 2026-10-24), #598 (debian base), #520/#513/#512/#510 (actions bumps), #467 (upload-pages-artifact), plus stale #423 (wen2so 2025-08, delete-design vs issue #417). No human feature PR is open — the trunk is quiet. Gate: human batching decision under the standing minimize-PR-count directive (the directive’s canonical location is not recorded in the scan).
Branch residue nam/getdesign-stream @ de306e1 fully merged (verified ancestor of 99f03f3) but un-deleted; nam/unified-model (2026-08-30) and nam/unified-clang-build (2025-12-27) each hold exactly 1 unmerged commit. Gate: cleanup triage.
Plan-index lag README.md (in-flight index, updated 2026-11-01 — re-read locally this session) references none of program-status-2026-11-01.md, m6-4-client-fanout.md, or getdesign-streaming-oom-plan.md; the OOM plan is absent from development entirely. Gate: next docs pass.

2.2 3D client

Item State / gate
development→main promotion of #67’s 3 commits The only near-term open work item; PR not yet opened (same shape as promotions #61/#44/#37). Gate: the secret-scan gate flaked on #62/#63/#65 and needed retrigger pushes (de88ed1, 3f9c227, 7187a8b) — rerun on “Failed to complete secret scan”; never read a cancelled run as red.
Local checkout unsynced On feat/getdesign-stream @ 6dd8592 whose upstream was deleted after merge (0 unique commits vs origin/development); local development stale at 0430d86 (3 behind); 3 stashes parked (messages devopt on dev/opt, diags on devopt/grpc-diag, changes on planpart2/render-components-finish — re-checked via git stash list this session) — contents not inspected; may hold NI24 diagnostics.
M6.4 post-#67 remainder Per m6-4-client-fanout.md §5 (a plan doc in this repo, read this session): after #67 the only remainder is the info-client stream fallback (§3.3 #1 / §4 step 5), plus server-side M6.6. Nothing further is owed in the 3D client itself.

2.3 Info client

2.4 Cross-cutting


3. Remaining work (backlog, by component)

Not execution plans — candidate work items, in rough priority order within each component.

3.1 Server

# Item State / gate
1 Promotion debt (the headline) — 91 commits on nam20485 never seen by development; staging frozen since 2026-09-14, main/release since 2026-09-08 development strict checks (ruleset 169324: 3 CMake legs, SBOM, Codacy, dependency-review, 1 approval); 169360 bases need --admin merges. The 1 approval has the same solo-maintainer question as 169360’s approvals — see §4 step 2
2 M6.6 residency budget + the REST big-design guard decision — the dispatch guard exists only on the gRPC path (OdbDesignServiceImpl.cpp:276 per predecessor §3.1-#3); REST serializes whatever the cache holds M6.6 cut from 99f03f3; guard decision rides the same PR
3 Auth track M2.1–M2.5 (stopgap Basic → Keycloak → JWT middleware/gRPC interceptor; issues #504/#505) · M3.2 WebSocket /events · M1.5 designs-dir watcher (README.md:14) planned, not started; after M6.6 + promotion; Keycloak provisioning
4 SI19 item 1.3 — arena allocation in streaming RPCs (README.md:17), descoped from #587 post-M6.6 perf work — more relevant now streaming is first-class
5 Argo CD GitOps execution (README.md:21; plan merged #568; zero Applications on cluster today) · IPC-2581 import (README.md:23-26; client guide describes a non-existent API) stable server line / future feature
6 Issue hygiene — 38 open: #599 closeable (fix merged via 6a8bfd1); semi-recent #517/#518 (Feb 2026), #504–#506 (Dec 2025); the rest 2025-and-older dormancy (PCB-viewer epics #440–#447, #473, #352, #248) triage sweep
7 Deferred by decision — M6.7 parse-footprint reduction (wire-dangerous, needs lazy re-hydration design); plan-index README refresh explicitly deferred / rides M6.6

3.2 3D client

# Item State / gate
1 development→main promotion (§2.2) PR open + secret-scan flake management
2 #64 — silkscreen mirrored on Turbot (top + bottom; live bug, open since 2026-10-24, zero comments) evidence committed at docs/plan/issues/mirrored_text/turbot_top_sst.png + turbot_bottom_ssb.png (commit 04919d0 — both files verified present locally; these document the bug as filed); same class as the #53 chirality fix; diagnosis additionally needs A/B comparison captures — blocked on capture/evidence tooling (NI24), see §4 step 8
3 NI24 — headless automation stall: 0%-CPU hang at frame=5/drawCalls=0 under Xvfb, reproduced with mock AND real backends; blocks every --auto-screenshot flow (NI23 sweeps, silk A/B for #64) diagnosis with dotnet-stack on the stalled process or WaitForAsync timeout logging (new-issues-remaining.md:5-21, re-read locally)
4 gRPC perf backlog R-2/R-5/R-6/R-7 (NI21 review; the opt doc §4 table, re-read this session): R-2 concurrent symbols prefetch (serial per layer, 7.3 s on rigidflex / 34 layers — cheapest measured win), R-5 streaming stall guard, R-6 concurrent layer streams (stream-wait 69% on rigidflex post-batch), R-7 GC re-check. R-1 and R-4 are done (2026-09-21); R-3 is a separate RequestLoadDesign proposal item, not part of this perf set baseline predates M6.4 — re-measure post-#67 before picking
5 Packaging & release plan unimplemented — draft 2026-09-10 (status “Draft for review”, verified locally); 0 git tags, no release branch owner decisions: private-repo Actions billing (macOS 10× multiplier) and the SharpEngine open-source license requiring the repo to be public
6 Org operational constraints (current) — unreliable artifact storage (coverage decoupled from cross-job artifacts in 887ace7; artifact-upload failures fail the job again in 7d4ce22, per the owner directive of 2026-10-23 cited verbatim in that commit’s message); secret-scan gate flakes; random run cancellations not in evidence this window (last 15 runs all success) operational, not code
7 #11 — hit testing & selection (2026-02-07) — largely delivered via #36/#38 + #19-era work verify-and-close or retitle

3.3 Info client

# Item State / gate
1 GetDesignStream client adoption — contract vendored (service.proto:35, :146-148) with zero hand-written client references (the only .cs hits are git-ignored, generated obj/ grpc codegen — §1.3); on the live #609 server the two big designs (200-40628, 350-41017) return a clean RESOURCE_EXHAUSTED that FetchDesignArchiveAsync swallows (DesignService.cs:429-434 — logs a warning, returns null) → design renders empty, no user-visible error product decision: mirror #67’s fallback+reassembler (per m6-4-client-fanout.md §4 — resolve map entries by iteration, never find()/at(); the #599 landmine is the address-layout-dependent protobuf-Map find() flake behind server issue #599, fixed server-side by a linear scan in 6a8bfd1 — the same ordering hazard for any client reassembler) or declare big designs out of scope; either way stop nulling genuine transport failures
2 Transport-truth correction — satisfied; no standing instance remains. Corrected in this document (callout below) and in m6-4-client-fanout.md §2 (its own correction note, line 27). A grep -rn REST-first sweep run this session across all three repos’ docs/ plus the info client’s src/ found no other occurrence — the only hits are the two correction records none — documentation-only
3 #10 obsolete — asks to implement REST JSON deserialization for components/nets: exactly what decision D5 deleted (DesignService.cs:27-31; docs/connectivity-task.md:16-18) maintainer closes as superseded
4 Phase epics #2–#8 predate the rebuild — re-triage, don’t execute literally: #5 done (#26); #3/#4/#6 substantially landed (12 tab views, BasicAuthService, connection state machine ConnectionService.cs:104-278); #7 pipe client wired into Components/Pins/Nets tab VMs (CrossProbeService.cs:15-26) but viewer-side counterpart unverified; #8 partial (#24 harness; no release/packaging) maintainer re-triage against post-rebuild reality
5 Proto refresh manual, no drift check (D7) — three hand-refreshes already (#23/#25/#27); silent server contract changes surface only at compile or runtime decision on where a drift check lives (client CI vs server publish)

Transport truth (code-cited; recorded per the info-client scan’s explicit requirement). gRPC GetDesign.FileModel is the single content source for all design reads — matrix, step headers, EDA data, symbols, components/nets/packages/parts, drill tools (src/OdbDesignInfoClient.Services/DesignService.cs:18-32 in nam20485/Odbdesign-info-client-india79-b; the fetch itself is DesignService.cs:413, re-read this session). REST survives only as discovery + health: /filemodels, /filemodels/{name}/steps, /healthz/* (Api/IOdbDesignRestApi.cs:23-47). The per-resource REST projections were deliberately deleted per decision D5. The 2026-11-01 lesson’s “REST-first; gRPC only for Connectivity” claim is false in code — and per the sweep above, no document still asserts it.

3.4 Cross-cutting


4. Action plan — one ordered sequence across all three projects

flowchart TD
    a1["1 3D: promote development to main\n3 commits - CI already green"]
    a2["2 S: promote nam20485 to development\n91 commits - 2-way merge"]
    a3["3 S: cut + execute M6.6\n+ REST guard decision"]
    a4["4 S: hygiene + cluster watch\n#599 - dependabot x7 - swaggerui"]
    a5["5 IC: GetDesignStream adoption\nfallback or descope"]
    a6["6 S: Argo CD GitOps execution\n0 Applications today"]
    a7["7 3D: fix NI24 headless stall"]
    a8["8 3D: diagnose #64 silkscreen"]
    a9["9 S: resume staging main release chain\n+ dev-based docker publish"]
    a10["10 3D: re-measure then R-2 perf\nR-5 R-6 R-7 follow"]
    a11["11 IC: proto-drift guard"]
    a12["12 IC: stale-issue + branch triage"]
    a13["13 3D: triage #11 - verify and close"]
    a14["14 S: mid-term auth M2 + WS M3.2\n+ SI19 arena item"]
    a15["15 3D: packaging release decisions"]
    a16["16 S: IPC-2581 import - future"]
    a2 -->|"stable line first"| a6
    a2 --> a9
    a3 -->|"perf items build on it"| a14
    a7 -->|"capture evidence"| a8

Discharged in this document: [IC] the transport-truth correction (§3.3 callout) — and the §3.3-#2 sweep confirms no other document still carries the claim, so no doc sweep remains open.

This document’s own landing — do not repeat the 59ed1cd mistake (§1.1): land this file on the nam20485 line (branch off nam20485, PR into nam20485), merged before step 2’s promotion, so it rides into development with the 91 instead of widening the two-way merge as a second straight-to-development commit. At scan time the server tree had 0 untracked files; this file is now that tree’s only untracked file (git status --porcelain → ?? docs/plan/program-status-2026-11-02.md, re-run this session).

  1. [3D] Promote development → main — promotion PR for #67’s 3 commits (same shape as #61/#44/#37). Waits on: nothing technical — the development-merge CI was fully green 2026-11-02T13:38Z; the operational gate is the secret-scan flake (rerun on “Failed to complete secret scan”; never read a cancelled run as red).
  2. [S] Promote nam20485 → development — 91 commits, now a two-way merge (development is 1 ahead via 59ed1cd); carries the connectivity + streaming epics and all CI repairs. Waits on: ruleset 169324 strict checks (3 CMake legs, SBOM, Codacy, dependency-review, 1 approval) — tip CI is already 4/4 green, so the risk is the gate, not the code. Open question to settle before the PR: 169324’s required 1 approval has the same solo-maintainer problem this doc flags on 169360 (the author cannot approve their own PR); the scan data does not record where that approval comes from or whether step 2 also needs --admin — 169360’s equivalent is resolved with --admin, 169324’s path is not recorded in this scan.
  3. [S] Cut nam/m66-residency-budget from 99f03f3 and execute M6.6 — resident-estimate budget (kResidencyFactor ≈ 5.0) + FixtureCapture --stream; take the REST big-design guard decision in the same PR (even if the answer is “document unsupported”). Waits on: nothing — the branch cut is the first action; lands on nam20485 and rides the next promotion cycle.
  4. [S] Hygiene + cluster-watch batch — close #599 (fix merged via 6a8bfd1 on the #609 line); disposition all seven open dependabot PRs, not just the two with recorded intent (#607 vcpkg / #598 debian base = fold into the next PR): #520/#513/#512/#510 (actions bumps) and #467 (upload-pages-artifact) need an explicit merge/close/ride call in the same batch — no disposition for those five is recorded in the scan; dismiss or restate stale #423; cheap triage sweep of the 38 open issues. Cluster watch: decide the SwaggerUI restart count — 60 restarts since 2026-08-20, last 3d18h before the probe, no cause recorded: investigate (pod events/logs for odbdesign-server-swaggerui-v1-dff4bfc8c-tgv8b) or explicitly waive, and record the decision here. Waits on: human action only.
  5. [IC] GetDesignStream adoption decision + implementation — mirror #67’s fallback+reassembler or declare big designs out of scope; either way stop swallowing genuine failures (DesignService.cs:429-434). Waits on: a product decision in the info-client repo — the technical gate (server endpoint live) is satisfied: the cluster runs the #609 build (today’s probe) and the predecessor recorded live acceptance.
  6. [S] Start Argo CD GitOps execution — Applications per argocd-deployment-plan.md (plan merged #568; zero Applications on the cluster today); retires the manual deploy.ps1/kubectl chain. Waits on: step 2 landing first is preferable (stable development line); deploy-branch decisions per the plan.
  7. [3D] Fix NI24 headless stall — dotnet-stack on the stalled process per the runbook (new-issues-remaining.md:20-21); inspect the 3 parked stashes first (may hold diagnostics — contents unverified). Waits on: a local diagnosis session; nothing upstream.
  8. [3D] Diagnose #64 Turbot mirrored silkscreen — start from the rev-4 numeric methodology in docs/plan/issues/mirrored_text/ANALYSIS.md. Waits on: fresh A/B capture evidence — the committed PNGs (04919d0) document the bug as filed; the diagnosis needs comparison captures (the “silk A/B” variants the --auto-screenshot flow would produce — exactly what NI24 blocks), or manual screenshots in the meantime.
  9. [S] Resume the frozen chain — staging (frozen 2026-09-14) → main → release (frozen 2026-09-08) + a development-based docker publish. Waits on: step 2 landing — per the scan, chain resumption gates on the promotion only. M6.6 is not a gate: it lands on nam20485 (step 3) after the chain resumes and rides the next nam20485→development promotion, so the resumed chain carries today’s 91 commits, not M6.6. 169360 bases need --admin merges (code-owner + last-push approval unsatisfiable solo).
  10. [3D] Re-measure the post-#67 load profile, then R-2 concurrent symbols prefetch; R-5/R-6/R-7 follow per what the re-measure shows. Waits on: the re-measurement — the existing numbers (symbols 7.3 s serial on rigidflex; stream-wait 69%) predate GetDesignStream.
  11. [IC] Add a proto-drift guard — diff the vendored protoc/ against the server repo’s, in CI. Waits on: a decision on where it lives (client CI vs server publish) — D7 currently records “no drift check exists or is planned”.
  12. [IC] Stale-issue + branch triage — #10 close-as-superseded; #2–#8 re-triage against the rebuilt client; delete merged/stale remote branches (feature/f2-conformance-harness, feature/f3-grpc-design-reads, feature/phase-3-parity, copilot/*, mn/droid-app, add-factory-workflows-*). Waits on: maintainer decisions only.
  13. [3D] Triage #11 (hit testing/selection) — largely delivered (#36/#38 + #19-era work); verify remaining gaps, then close or retitle. Waits on: nothing.
  14. [S] Mid-term tracks — auth M2.1–M2.5 (Keycloak first; issues #504/#505), M3.2 WebSocket /events, SI19 item 1.3 arena allocation in streaming RPCs. Waits on: M6.6 + promotion; Keycloak provisioning.
  15. [3D] Decide then execute the packaging/release track — the plan is fully specified but nothing landed (0 tags, no release branch). Waits on: owner decisions — private-repo Actions billing (macOS 10×) and the SharpEngine open-source license requiring a public repo.
  16. [S] IPC-2581 import — future feature; three planning docs exist, and the client migration guide describes an API that does not exist yet. Waits on: upstream prioritization.

Appendix — verified state snapshot (2026-11-02)

What Value
Server branches development 59ed1cd (local = origin; clean at scan time with 0 untracked files — this document is now that tree’s sole untracked file, git status --porcelain re-run this session) · nam20485 99f03f3 (91 ahead / 1 behind development) · staging 9748709 (51 ahead / 8 behind development; frozen 2026-09-14) · main 6074592 (2 ahead / 56 behind staging; frozen 2026-09-08) · release 33448b5 (2 ahead / 0 behind main)
3D branches development e455dd7 · main 0430d86 (main..development = 3: 2299e6b, 6dd8592, e455dd7; reverse = 0) · local on feat/getdesign-stream 6dd8592 (upstream deleted post-merge), clean, 3 stashes (messages devopt on dev/opt, diags on devopt/grpc-diag, changes on planpart2/render-components-finish); local development stale 3 behind; 0 tags
Info branches main c98a179 · development 72f700d · nam20485 8679ac8 — nam20485 ⊆ development ⊆ main (reverse drift = 4 promotion merge commits only); local nam20485 clean; stale local main (41 behind) and development (33 behind)
Server open PRs / issues 8 PRs = 7 dependabot (#607, #598, #520, #513, #512, #510, #467) + stale #423 · 38 issues, only #599 recently updated (2026-10-30; closeable — fix merged via 6a8bfd1)
3D open PRs / issues 0 PRs · 2 issues — #64 (mirrored silkscreen, 2026-10-24), #11 (hit testing, 2026-02-07)
Info open PRs / issues 0 PRs · 8 issues — #2–#8 plus #10 (Feb-2026 phase epics, stale vs the rebuild; #9 is not among them)
Cluster (read-only probe) 18/18 deployments READY 1/1 across 6 namespaces (argocd, default, kube-system, monitoring, portainer, trivy-system). odbdesign-server-v1 1/1, imageID ghcr.io/nam20485/odbdesign@sha256:29a4501d0562…, pod odbdesign-server-v1-5cc87f5ddc-mxnmb start 2026-11-02T00:21:22Z, 0 restarts. odbdesign-server-swaggerui-v1 1/1, imageID …swaggerui@sha256:4cf342ddb938…, pod start 2026-08-20T05:10:23Z, 60 restarts (last 3d18h before probe; cause not investigated — decision owed, §4 step 4). Argo CD: 6 deployments 1/1 (56d old), 0 Application CRs cluster-wide (kubectl get applications.argoproj.io -A → No resources found)
Not verified in this scan Live streaming behavior of the running server (acceptance quoted from predecessor line 8); the identity of the dependabot PR that closed (8→7); contents of getdesign-streaming-oom-plan.md (nam20485-only); the 3 local 3D stashes; runtime/UX state of the info client’s 12 tab views; the cause of the SwaggerUI restarts; whether 169324’s required 1 approval is satisfiable by a solo maintainer or needs --admin (§4 step 2); no builds or tests executed — all scans read-only