19 KiB
CI/CD Monitor Agent — Persistent Memory
Persistent diary cross-session. Auto-injected first ~200 lines at spawn (L1 HOT). Update BEFORE every stop. Tiered Memory v1: L1 HOT soft-cap ~30KB · L2
archive/on-demand · L3 RAGsearch_memoryjust-in-time. Keep entry ≤ 1.5K chars (gotcha #53). Full verbatim run history pre-S40 → gitd2f52ba+archive/2026-05-{runs,q2,q3,q4}.md+archive/2026-06.md+archive/2026-07.md. 🗺️ Lookup map:archive/_INDEX.md— 1 dòng/bản-ghi + con-trỏ substring (sha-keyed, Ctrl-F fallback); đọc verbatim +2026-0{5,6}.gist.mdtheo nhu cầu. See 🔄 Curate-trigger log for archive-move history.
🎯 Role baseline
Read-only CI/CD + post-deploy verifier SOLUTION_ERP. Polls Gitea Actions API, verifies test gate + deploy ship + prod health. Tools: Read, Grep, Glob, Bash, WebFetch + 5 RAG MCP. Output: PASS/FAIL + evidence <500 words. Skills: iis-deploy-runbook + dependency-audit-erp + ef-core-migration. Spawn ~150K — trade-off catch fail tự động.
🚨 Recurring CI/CD bug patterns (catch priority)
- #39 act_runner github.com TCP timeout — run hang "Set up job" 21s. Log
dial tcp github.com:443 i/o timeout. Fix: manual checkout bypass hardcoded.gitea/workflows/deploy.yml. KHÔNG revert. - #40 npm cache
tsc not found—build_fe_adminfail postcache: npm. DISABLED rolled backa21790d. KHÔNG re-enable. - #41 paths-ignore docs-only skip — code commit không trigger CI? Check
git diff --name-only HEAD~1 HEADvspaths-ignore7-entry từ D10 S1118aa3869(#495):['docs/**','**/*.md','.claude/skills/**','.claude/agent-memory/**','.claude/workflows/runs/**','.gitignore','scripts/**.md']. Gitea evaluates push range — ≥1 commit có non-ignored file → toàn range build. ⚠️ AS-6 BASELINE SHIFT: từ #495, memory-push chứamemory-budget.json/agent-memory-file/run-trace HẾT trigger CI (pattern #490/#492 DEAD). - #25 IIS WebSocket —
notification-hub/negotiate401/404 prod. Fix: WebSocket module enableweb.configsite api (skilliis-deploy-runbook). - #48 SQLite tie-break —
OrderByDescending(CreatedAt).First()pick wrong khi 2+.Add()cùng frozen-clock. Fix: discriminator filter BEFORE OrderBy. - Bundle hash unchanged = ship FAIL — push+action success nhưng prod không đổi. Verify via INDEX.HTML ref (
curl -s https://admin.solutions.com.vn/ | grep -oE '/assets/index-[A-Za-z0-9_-]+\.(js|css)'), NOT by GETting a hash-named asset directly. ⚠️ verify MUST sau status=success (Run #242 false-positive: check khi "running" → stale hash).- 🔴 #69 (S72 Run #312) — FE bundle hash NON-DETERMINISTIC per rebuild.
deploy.ymlRemove-Item fe-*\* -Exclude web.config+Copy-Item dist\*runs UNCONDITIONALLY every run (path-filter gates whole-workflow trigger, NOT per-step). Identical FE source ⟹ DIFFERENT hash each deploy (Vite/rolldown non-reproducible chunk-id). PROVEN: governance-only18fced6(0 files in fe-*/src) rotated BOTH bundles. ⟹ "BE-only/governance ⟹ bundle frozen" is FALSE; rotation EXPECTED every deploy. To detect REAL FE change, difffe-*/srcin commit/range, NOT hash delta. - ⚠️ SPA-fallback 200 trap (S72): server rewrites
/*→/index.htmlso GET/assets/index-<ANYTHING>.jsreturns 200 even for fake hash (controlZZdoesnotexist0.js→200, size~919b text/html). Old-hash-still-200 MEANINGLESS. RELIABLE signal = parse index.html<script src>, GET that exact path + checksize_downloadLARGE (real ~1.6MB vs fake ~919b) +Last-Modifiedin deploy window + 2nd-fetch byte-stable (no mid-deploy transient). - 🟢 #69-REFINE (S83 Run #333) — hash can stay IDENTICAL across deploy when fe-*/src 0-change. #69 said "rotates EVERY deploy" — NOT absolute. Run #333 (BE+tests+governance, ZERO
fe-*/srcdiff) shipped FE (Last-Modified 17:10:35/17:11:27 BOTH in run-window 17:06→17:11) but hash UNCHANGED (adminCsJetgZH/userBVS0ApIm= #330 baseline). ⟹ identical Vite source ⟹ byte-identical output ⟹ same content-hash. Rolldown chunk-id apparently MORE deterministic than #69 feared (or only rotates on real src delta). MOST RELIABLE ship-proof =Last-Modifiedheader on the real bundle file IN deploy-window, NOT hash-delta (hash-delta has false-NEGATIVE when src unchanged + false-POSITIVE risk noted in #69). Cross-refgit diff fe-*/srcscope: 0-FE-change ⟹ hash EITHER same (#333/#485-487) OR rotated (#312/#489 S107) — both benign; ship-proof = fresh LM-in-window + real-size.
- 🔴 #69 (S72 Run #312) — FE bundle hash NON-DETERMINISTIC per rebuild.
- Migration drift prod vs repo — compare
ls .../Persistence/Migrations/*.csvssqlcmd __EFMigrationsHistory. Fix: checkProgram.cs/DbInitializerapp.MigrateDatabase()+ app pool recycle. Mig-applied proof =__EFMigrationsHistorytop==repo-HEAD; stuck-old ⟹ pool didn't recycle ⟹ FAIL even if status=success+bundle-rotated.
📋 5-stage checklist (EVERY run)
- Stage 0 RAG infra:
Get-Service QdrantRunning +http://localhost:6333/healthz. Collectionproj_solution_erp. - Stage 1 Push+filter:
git log -1 --format='%H %s'+git log origin/main..HEADempty + diff vs paths-ignore (docs-only → SKIPPED-DOCS return). - Stage 2 Gitea poll (max 10 iter × 60s): API
.../actions/tasks?limit=5(NOT/runs404). Matchhead_sha(NOT run_number — cancelled runs shift numbering). ⚠️updated_atstale ~2min (gotcha #46) → cross-check VPS mtime. Foreground-sleep BLOCKED Windows → Monitor-style busy-wait-curl-spin ~30s/iter. - Stage 3 Test gate: baseline 440 PASS (45 Domain + 395 Infra; S96 #480
0c2f9f6— 436 + newPeSuggestedPriceSetterAuthzTests[price-lock-after-submit authz]; prior S96 #479 436, S96/S95 434, S90 #465 421). CI runs both proj BEFORE build/deploy → status=success ⟹ test gate passed (tasksterminal=status:success,conclusionNOT populated — trust success NOT log-numeric). Local grep undercounts (Theory/InlineData). Phase 9 UAT skip OK. - Stage 4 Post-deploy (if SUCCESS): auth login bearer (admin + nv.test gotcha #44; token=
accessTokenroute/api/auth/login) → 3-5 endpoint smoke 2XX (incl new — NEW-endpoint bodyless-POST returns 411 IIS Length-Required pre-auth NOT 404; re-probe WITH body → 401 confirms wired) → FE bundle hash 2 app (per #69 = real-size-vs-fake not hash-delta) → SignalR negotiate (gotcha #25) → EF mig prod==repo.- Stage 4.6 (S29 CRITICAL): sqlcmd seed sample verify post-deploy (NOT chỉ schema). Seed=runtime-row-insert (no history/tables advance) → verify via DB-query (e.g. Permissions count), NOT bundle-frozen-alone.
- Discovery: ASP.NET 10 record enum needs numeric input unless
JsonStringEnumConverter(SOL has NO converter → FE sends numeric). sqlcmd ssh Windows-auth\\\\SQLEXPRESS4-backslash. INFRASTRUCTURE seed MUST run (NOT insideif(!demoSeedDisabled)); DEMO seed gated → gotcha #51.
- Stage 5 Report PASS/FAIL + evidence + MEMORY update.
⚠️ Anti-patterns (DO NOT)
- ❌ Push fix code — READ only, escalate em main · 2. ❌ Speculate fail without log · 3. ❌ Skip post-deploy bundle hash (biggest catch) · 4. ❌ Skip MEMORY · 5. ❌ Poll forever (max 10 iter) · 6. ❌ Auto-rollback (escalate + recommend) · 7. ❌ Verify docs-only (SKIPPED-DOCS return ngay)
🧠 SOLUTION_ERP CI/CD essentials (S80 verified — re-ground từ docs/STATUS.md canonical)
- Gitea:
git.baocaogiaoduc.vn/vietreport-admin/solution-erp· workflow.gitea/workflows/deploy.yml· paths-ignore 7-entry (D10 S111 #495):['docs/**','**/*.md','.claude/skills/**','.claude/agent-memory/**','.claude/workflows/runs/**','.gitignore','scripts/**.md']. Anon API works (public repo) when GITEA_TOKEN absent. - Prod: api/admin/eoffice
.solutions.com.vn· SSHssh vietreport-vps(Administrator, id_ed25519) · IIS phys paths: APIC:\inetpub\solution-erp\api· admin\fe-admin· user\fe-user. DB.\SQLEXPRESS/SolutionErp/vrappSQL-auth. Conn key =ConnectionStrings.Default(NOTDefaultConnection). DB pw =buKL3TGBkD0wDDbYVw65QeX9(User Id=vrapp; ⚠️ Password field has NOvrapp/prefix — S89 login-failed when prefix included; read verbatim from prodappsettings.Production.jsonConnectionStrings.Default when$env:PROD_DB_PASSWORDempty). ⚠️ skill-doc pathC:\inetpub\apps\SolutionErp\ApiSTALE.- SSH→PS quoting (S42): nested bash→ssh→powershell mangles
$var/\"→ useiconv UTF-16LE | base64→powershell -EncodedCommand $B64; OR write ps1-file +powershell -File $(cygpath -w). sqlcmd over SSH direct-W -h -1pw-inline. sys-catalog string-concat →COLLATE DATABASE_DEFAULT(collation conflict).
- SSH→PS quoting (S42): nested bash→ssh→powershell mangles
- Tests baseline: 458 PASS (45 Domain + 413 Infra · 0 fail/skip; D10 S112 Run #496 sha
e100ef0— 440→458 incl +10SupplierExcelImportServiceTests+PeSubmitGuardAndBypassmods; prior 440 [S96 #480 45D+395IPeSuggestedPriceSetterAuthzTests], 436, 434, 421). Phase 9 UAT skip per chunk OK. 🔑 Anon log endpointhttps://git.baocaogiaoduc.vn/vietreport-admin/solution-erp/actions/runs/{run_number}/jobs/0/logs(HTTP 200 ~18KB) EXPOSES per-projectPassed! - Failed: 0, Passed: N, Skipped: 0— concrete test count now available, NOT just status=success (the/api/v1/.../tasks/{id}/logsvariant 404s; use web-UI runs/{run_number}/jobs/0/logs). - Mig latest repo: Mig 63
20260712070117_AddSupplierImportSourceFields(D10 S112 #496e100ef0; 2 AddColumn nullable onSuppliers[SourceUpdatedAtdatetime2 +SourceUpdatedBynvarchar(200)] for Excel-import Phase B provenance; NO CreateTable ⟹ sys.tables stays 89 — VERIFIED-APPLIED-PROD history-top==AddSupplierImportSourceFields+ both-cols-present=2 + sys.tables=89). Prev Mig 62ExpandSupplierFields(S104 #488, AddColumn ×18 Supplier 9→27 + un-cram 4-NCC Note). Prev Mig 61BackfillEndedByLevelFinalizeFromChangelog(S97, data-only). Pathsrc/Backend/SolutionErp.Infrastructure/Persistence/Migrations/. Prod checksqlcmd __EFMigrationsHistory ORDER BY MigrationId DESC TOP 5. ⚠️sys.tables(is_ms_shipped=0) = 89 (S104 LIVE-verified post-#488 — AddColumn/data-only migs don't change count → commit touches no new table ⟹ 89 correct don't FAIL. Cross-ref COMMIT scope vs ambient count). enum-additive (e.g.ApprovalAttachment=5) = int-stored NO-mig → history NOT-advance + emptygit diff -- '*Migrations*'. - Bundle hash live (D10 S112 #496
e100ef02026-07-12 — REAL-FE-SHIP rotate, breaks FROZEN×6): admin jsCixlt_MA/cssDs-E5neu· user jsBhBOG1dn/cssBeb0Bb1h— BOTH js+css rotated (from #490-#495 adminBVyH-FY7/Bbtesa6B· userEDsZebvn/BmuPz_Dy) = expected, realfe-*/srcchange (SupplierImportDialog.tsx NEW both apps). Ship-proof DEFINITIVE: real-size admin 1,652,140b / user 1,558,489b (vs 919b SPA-trap) +application/javascript+ LM 09:18:32/09:19:25 GMT (=16:18/16:19+07 in deploy-window 16:14:33→16:19:58) + content-marker"Database NCC"count=1 in BOTH bundles. ⚠️ id ≠ run_number; head_sha-keyed. ⚠️ PER #69 TWO-MODE: hash MAY rotate (real src-diff #488/#496, HOẶC 0-fe-src #489/#312) OR freeze (0 fe-src #485-487) — SNAPSHOT per-run, NOT baseline; ALWAYS re-snapshot LIVE pre-deploy. Real FE ship = difffe-*/srcin range + real-size-vs-fake + Last-Modified-in-window + content-marker-grep, NOT hash delta alone. - Bearer: admin
admin@solutions.com.vn/Admin@123456(full) · UATnv.test@solutions.com.vn/TestUser@123456(Drafter CCM, gotcha #44 check).
🔑 Critical config (flag commit nếu tái xuất)
Node CI 20.x (feedback_node_cicd) · MediatR 12.4.1 (gotcha #1, flag Version="14) · Swashbuckle 6.9.0 (gotcha #2) · act_runner manual checkout (#39) · npm cache DISABLED (#40, flag cache: npm)
🎯 Per-NV admin opt-in wire — 10-point checklist (cumulative S22→S23)
Cross-ref feedback_per_nv_permission_scope. Per-NV/per-Level refactor MUST verify: 1 Domain field · 2 EF HasDefaultValue(false) · 3 Mig 3-file · 4 Service read · 5 Domain+App DTO mirror · 6 Designer FE checkbox · 7 AwLevelDto+ToDto · 8 CreateAwLevelInput+Update mutation · 9 Lookup discrimination (FirstOrDefault ADD ApproverUserId==actorId + admin fallback) · 10 Controller body record count == Command record count. Bug latency 2-3 days prod silent khi miss 9-10. Scan grep -n "FirstOrDefault.*Order.*==" *.cs after OR-of-N refactor.
📊 Run stats baseline
BE (test+build) ~90s · FE × 2 ~60s/app · deploy ~30s · total ~3min code / 0s docs-only. >5min → escalate. Recent runs all ~4m40-5m.
📅 Recent runs (compressed — verbatim → archive via _INDEX.md)
- 🟢 S112 #496 (rn 382, id 496)
e100ef0GO 5m25s (16:14:33→16:19:58+07 2026-07-12) — PE workflow (a829d0dcode-only A1-creator-sign/EDGE-5-terminal-lock/opinion-history) + Supplier Excel-import Phase B + Mig 63 (22-file: FE ×4 SupplierImportDialog both apps, BE ×13 incl Mig63 3-file, tests): status=success · test gate CONCRETE fromruns/382/jobs/0/logs(anon 200): DomainPassed! Failed 0, Passed 45+ InfraPassed! Failed 0, Passed 413= 458/0-fail/0-skip (matches parent baseline exact; +10 SupplierExcelImportServiceTests absorbed) · Mig 63AddSupplierImportSourceFieldsAPPLIED-PROD (history-top==name +SourceUpdatedAt+SourceUpdatedBycols-present=2 + sys.tables=89 frozen) · FE REAL-SHIP both apps admin jsBVyH-FY7→Cixlt_MA/css→Ds-E5neu· user jsEDsZebvn→BhBOG1dn/css→Beb0Bb1h(all 4 rotated, real fe-src) ship-proof= real-size 1,652,140/1,558,489b +application/javascript+ LM 09:18:32/09:19:25 GMT in-window +"Database NCC"marker count=1 both · import routes WIRED POST/import/preview+body→415 (multipart binder matched) ·/import/confirm+body→401 (auth-gated) · bogus+body→404 control · smoke api-live/ready + admin/eoffice roots + authed suppliers/contracts/PE/menus ALL 200 (PE workflow change healthy). 🔑 411-refinement: bodyless-POST → 411 for ANY route incl bogus (IIS Content-Length pre-routing) ⟹ 411≠route-exists; MUST re-probe WITH body → 401/415 (matched) vs 404 (missing). - 🟢 S111 #495 (rn 381, id 495)
8aa3869GO 5m11s (11:45:21→11:50:32+07 2026-07-12, iter6) — H22 WAL adopt 30-file +767 (0 src/BE/FE/mig; non-ignored: settings.json + wal-flush.ps1 + hmw.js + deploy.yml + governance-detectors.ps1 + .gitattributes → CI-run ĐÚNG thiết kế): status=success (gate 440) · deploy.yml TỰ-ĐỔI trong chính commit (D10 +2 paths-ignore.claude/agent-memory/**+.claude/workflows/runs/**) — run created+parse+chạy trọn 5m11s = YAML hợp lệ, workflow-self-change PASS · Mig FROZEN 62 history-topExpandSupplierFields+ sys.tables=89 ✓ · smoke 3×200 (health/ready + 2 FE roots) · bundle FROZEN ×6-streak (adminBVyH-FY7/Bbtesa6B· userEDsZebvn/BmuPz_Dy= #490→#494) + LM-in-window 11:49:07/11:50:00+07 + size 1,644,931/1,551,280b byte-identical #494 vs fake-919b = ship-proof. 🔑 AS-6 BASELINE SHIFT từ #495: paths-ignore 7-entry — memory-budget.json/agent-memory/run-trace push HẾT trigger (trigger-pattern #490/#492 DEAD); phân loại trigger giờ soi file ngoài 7-glob. - (S107-closeout→S110-spot #490-#494 GOVERNANCE/MEMORY-push GO ×5 [
d74f67f·69957fd·f5a5c3e·6ee60cd·ac3f122— trigger #41 non-.md-in-push: budget.json ×2 + hmw.js ×3, pattern DEAD từ D10 #495] — bundle FROZEN-streak ×5 sau ROTATE #489 [#69 TWO-MODE; LM-in-window + real-size 1,644,931/1,551,280b = signal duy nhất] · Mig 62 + sys.tables 89 frozen · gate 440 · smoke 3×200 ALL · 🔑 #494 spot-detection: LM-prod-newer-than-last-verified-window = missed-run detector · full verbatim →archive/2026-07.mdvia_INDEX) - (S107 #489
21387cdGOVERNANCE-ONLY (+office-document 12th sub) GO 5m33s — JS ROTATED-benignChSNAJBK→BVyH-FY7/Bn2w1znH→EDsZebvnCSS-frozen, 🔑 #69 TWO-MODE CONFIRMED (0-fe-src ROTATE #312/#489 hoặc FROZEN #333/#485-487 — ship-proof = LM-in-window + real-size, hash-delta phụ) · S104 #488778fc98FULL-STACK Supplier 9→27 + Mig 62ExpandSupplierFieldsAPPLIED-PROD — AddColumn×18-exact + sys.tables-89-invariant + un-cram 4-NCC Note=NULL/Pkg/Bank/LegalRep-populated-EXACT + DTO-18-new-keys-live + JS-rotate-real-src + 🔑 FULL-STACK-SCHEMA-ADD+DATA-UN-CRAM PASS-pattern — full verbatim →archive/2026-07.mdvia_INDEX) - (S103 #487 + S100 #486 + S98 #485 GOVERNANCE/SCRIPTS-ONLY GO ×3 [.ps1/.json/.md-mix non-ignored → #41] — bundle FROZEN #69-REFINE ×4, GOVERNANCE-PASS stable [#476/#485-487] · 🔧 python
time.sleepforeground-OK (chỉ bash-sleep-block) · full verbatim →archive/2026-07.md(+#485 gitc112be9) via_INDEX) - (S96+S97 FINAL-REVIEW #478-#484 11-commit ALL-HEALTHY whole-session-GO [backfill A/005=1/A/008=1/A/010=0-exact + 88-excl/89-incl reconcile; #69 rotation≠ship-signal] · S97 #484
fb2dfa9DATA-FIX Mig61 🔑 DATA-FIX-BACKFILL PASS = CI + Mig-top==HEAD + target-rows-flip-EXACT + control-untouched + derived-API-field-FLIP-live — verbatim →archive/2026-07.mdvia_INDEX) - (S96-S97 endsBeforeCeo-saga + PE-fixes #479-#484 all PASS 440 — key: #482 Mig60 CROSS-STACK [🔴 EF-TERNARY-SQL-translate-LIVE list-200-NOT-500 · #77 field-PRESENT] · #481/#483 FE-only [#77 new-str-PRESENT/old-ABSENT] · #480 baseline-440 +PeSuggestedPriceSetterAuthzTests · JS-rotate/CSS-frozen · verbatim → git +
2026-06.mdvia_INDEX) - (S90→S96 #465·#470·#472·#473·#476·#477 all PASS — key: #477 🔑 DATA-SEED-PASS = pre/post-DB-row-EXACT-delta + DLL-mtime + w3wp-recycle + Mig-frozen + API-payload-contains-row [⚠️ sqlcmd
-Wdiacritic-LIKE false-NEG →_-wildcard; base64-EncodedCmd-FAIL →.ps1-File] · #472 ⚠️ PROD Identity tables NON-prefixedRolesNOTAspNetRoles· #470 GO-LIVE 🔑 no-resurrect = pre+post-DB-count-diff + dll-in-window + w3wp≥run-window [integrated-auth-E] · #476 GOVERNANCE-PASS baseline · verbatim → git +2026-06.mdvia_INDEX) - (S69b→S89 runs #307·#313·#315·#318·#330·#333·#342-#348 all PASS — key: Mig54-59 applied-prod · #345 PATCH-401+7-authz · #344 backfill-winners=13 · #343 BE-only ship-proof=Infra.dll-mtime · #342 INVERTED-PASS-FROZEN [#73 TZ] · #346 FE-only js+css-BOTH-rotate · #348 411→401 endpoint-wired · full verbatim → git +
2026-06.mdvia_INDEX) - Older runs (S83 #325 FE-budget-indent ← S77 #329 banner ← S75 #301 ← S62 #286 ← … → S29 #232) full verbatim → git
d2f52ba+archive/2026-06.md(incl #291 06-16 FAIL forensic [gotcha #65]); pre-S38 →archive/2026-05-{runs,q2,q3,q4}.md.
🔄 Curate trigger
-
~30KB → archive recent runs → L2
archive/<period>.md(byte-exact append) +_INDEX.mdsubstring pointer. Dup failure patterns → merge. Stale >3mo → remove. - Last curate: 2026-07-12 S111 (self, PostToolUse 20.5→~17KB): #490-#494 verbatim →
archive/2026-07.md+_INDEX+5 pointers → 1 merged stub; nén S89-stub + curate-log. - Prev: S107 07-10 self 20.4→16.9 (#489/#488→archive) · S104 07-06 self 20.5→17.9 · S98 07-02 em-main 19.9→17.06 (
2026-07.mdNEW) · S80 06-20 em-main 38.8→16 · S70 06-17 Harness-9 65.2→23.2 (_INDEX/gist gen:1) · earlier S66 · S40.