fix(sync): use CW watermark incremental path for critical tables

User rows have a FK constraint to CwMember (User_cwMemberId_fkey). Syncing
Users first caused all 140 User upserts to fail since the CwMember table was
empty. This cascade failure then caused all Opportunity upserts to fail because
Opportunity.primarySalesRepId is FK-constrained to User.cwIdentifier.

Fix: reorder steps so CW Members syncs first, then Users.

.github/workflows/deploy.yaml

@@ -130,6 +130,17 @@ jobs:
             ghcr.io/horizonstacksoftware/optima-api-migrate:latest
             ghcr.io/horizonstacksoftware/optima-api-migrate:${{ github.event.release.tag_name }}
 
+      - name: Build and push the dalpuri sync image
+        uses: docker/build-push-action@v6
+        with:
+          context: .
+          file: api/Dockerfile
+          push: true
+          target: dalpuri-sync
+          tags: |
+            ghcr.io/horizonstacksoftware/optima-dalpuri-sync:latest
+            ghcr.io/horizonstacksoftware/optima-dalpuri-sync:${{ github.event.release.tag_name }}
+
   build-worker:
     name: Build - Worker
     needs: [test-api, test-dalpuri, test-ui]
@@ -220,9 +231,10 @@ jobs:
         run: bun install --frozen-lockfile
 
       - name: Rebuild native modules
-        run: npm rebuild
+        run: npm rebuild --ignore-scripts
         env:
           HUSKY: "0"
+          HUSKY_SKIP_INSTALL: "1"
 
       - name: Build macOS distributables
         run: bun run make:macos
@@ -261,9 +273,10 @@ jobs:
         run: bun install --frozen-lockfile
 
       - name: Rebuild native modules
-        run: npm rebuild
+        run: npm rebuild --ignore-scripts
         env:
           HUSKY: "0"
+          HUSKY_SKIP_INSTALL: "1"
 
       - name: Build Windows distributables
         run: bun run make -- --platform win32
@@ -276,6 +289,55 @@ jobs:
           files: |
             ui/out/make/**/*.exe
 
+  # Runs a full CW → API data sync as a Kubernetes Job (the CW MSSQL and
+  # API Postgres addresses are internal to the cluster and unreachable from
+  # GitHub-hosted runners).  Waits for both images to be built first and
+  # must succeed before either the API or worker deploys.
+  sync-cw-to-api:
+    name: Sync - CW to API
+    needs: [migrate-api, build-worker]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Set the Kubernetes context
+        uses: azure/k8s-set-context@v2
+        with:
+          method: kubeconfig
+          kubeconfig: ${{ secrets.KUBECONFIG }}
+
+      - name: Checkout source code
+        uses: actions/checkout@v4
+
+      - name: Delete previous sync job if exists
+        run: kubectl delete job -n optima -l app=dalpuri-sync --ignore-not-found
+
+      - name: Apply sync job
+        run: |
+          TAG=${{ github.event.release.tag_name }}
+          sed "s/RELEASE_TAG/${TAG}/g" dalpuri/kubernetes/sync-job.yaml | kubectl apply -f -
+
+      - name: Wait for sync to complete
+        run: |
+          TAG=${{ github.event.release.tag_name }}
+          JOB="job/dalpuri-sync-${TAG}"
+
+          kubectl wait --for=condition=complete --timeout=7200s -n optima "$JOB" &
+          WAIT_COMPLETE=$!
+          kubectl wait --for=condition=failed --timeout=7200s -n optima "$JOB" &
+          WAIT_FAILED=$!
+
+          wait -n $WAIT_COMPLETE $WAIT_FAILED
+
+          echo "--- Sync job logs ---"
+          kubectl logs -n optima "$JOB" --tail=500 || true
+
+          if kubectl get -n optima "$JOB" -o jsonpath='{.status.conditions[?(@.type=="Complete")].status}' | grep -q "True"; then
+            echo "Sync completed successfully."
+            exit 0
+          else
+            echo "Sync FAILED."
+            exit 1
+          fi
+
   # ==========================================================================
   # Deploy jobs
   # ==========================================================================
@@ -332,7 +394,7 @@ jobs:
 
   deploy-api:
     name: Deploy - API
-    needs: [migrate-api]
+    needs: [migrate-api, sync-cw-to-api]
     runs-on: ubuntu-latest
     steps:
       - name: Set the Kubernetes context
@@ -402,7 +464,7 @@ jobs:
 
   deploy-worker:
     name: Deploy - Worker
-    needs: [build-worker]
+    needs: [build-worker, sync-cw-to-api]
     runs-on: ubuntu-latest
     steps:
       - name: Set the Kubernetes context

api/Dockerfile

@@ -90,6 +90,9 @@ COPY --from=build /app/dalpuri/generated/ ./dalpuri/generated/
 # Copy production node_modules (Prisma adapter needs native bindings)
 COPY --from=deps /app/node_modules/ ./node_modules/
 
+# Ensure pdfmake Roboto fonts are present at runtime for PDF generation.
+COPY --from=build /app/api/node_modules/pdfmake/build/fonts/ ./node_modules/pdfmake/build/fonts/
+
 ENV NODE_ENV=production
 
 # ---- Stage 4: API server runtime image ----
@@ -133,4 +136,34 @@ COPY api/prisma.config.ts ./api/prisma.config.ts
 RUN chmod +x /app/api/prisma/migrate-entrypoint.sh
 
 WORKDIR /app/api
-CMD ["sh", "prisma/migrate-entrypoint.sh"]
+CMD ["sh", "prisma/migrate-entrypoint.sh"]
+
+# ---- Stage 7: Dalpuri CW-to-API sync runner ----
+FROM oven/bun:1.3.11 AS dalpuri-sync
+
+WORKDIR /app
+
+COPY package.json bun.lock ./
+COPY api/package.json ./api/package.json
+COPY dalpuri/package.json ./dalpuri/package.json
+COPY ui/package.json ./ui/package.json
+COPY patches ./patches
+
+RUN bun install --frozen-lockfile
+
+COPY dalpuri/src/ ./dalpuri/src/
+COPY dalpuri/prisma/ ./dalpuri/prisma/
+COPY dalpuri/prisma.config.ts ./dalpuri/prisma.config.ts
+
+COPY api/prisma/ ./api/prisma/
+COPY api/prisma.config.ts ./api/prisma.config.ts
+
+WORKDIR /app/dalpuri
+RUN DATABASE_URL="sqlserver://localhost:1433;database=dummy;user=dummy;password=dummy;trustServerCertificate=true" \
+    bunx prisma generate
+
+WORKDIR /app/api
+RUN DATABASE_URL="postgresql://dummy:dummy@localhost:5432/dummy" bunx prisma generate
+
+WORKDIR /app/dalpuri
+CMD ["bun", "run", "src/sync.ts"]

api/kubernetes/worker-deployment.yaml

@@ -20,6 +20,11 @@ spec:
           env:
             - name: MANAGER_SOCKET_URL
               value: "http://optima-api.optima.svc.cluster.local:8671"
+            - name: API_DATABASE_URL
+              valueFrom:
+                secretKeyRef:
+                  name: api-env-secret
+                  key: DATABASE_URL
           envFrom:
             - secretRef:
                 name: api-env-secret

api/prisma/migrate-entrypoint.sh

@@ -26,8 +26,11 @@ while [ $ATTEMPT -lt $MAX_RETRIES ]; do
   # P3009: a previously-failed migration is blocking deploy.
   # The error message contains the migration name in backticks:
   #   The `20260402000000_fix_severity_typo` migration started at ... failed
-  if echo "$DEPLOY_OUTPUT" | grep -q "P3009"; then
-    FAILED=$(echo "$DEPLOY_OUTPUT" | grep -oE '\`[0-9]{14}(_[a-zA-Z_]+)?\`' | tr -d '\`' | head -1)
+  # Strip ANSI escape codes first (Prisma may colorize output even without TTY),
+  # then use a simple backtick-content regex rather than a rigid format match.
+  CLEAN_OUTPUT=$(printf '%s\n' "$DEPLOY_OUTPUT" | sed 's/\x1b\[[0-9;]*[mGKHFJr]//g')
+  if printf '%s\n' "$CLEAN_OUTPUT" | grep -q "P3009"; then
+    FAILED=$(printf '%s\n' "$CLEAN_OUTPUT" | grep -o '`[^`]*`' | grep '[0-9]' | tr -d '`' | head -1)
     if [ -n "$FAILED" ]; then
       echo "[migrate] Resolving failed migration as rolled-back: $FAILED"
       RESOLVE_OUTPUT=""

api/prisma/migrations/20260408100000_sync_schema_from_db_push/migration.sql

@@ -210,6 +210,8 @@ CREATE UNIQUE INDEX IF NOT EXISTS "CatalogItem_id_key" ON "CatalogItem"("id");
 
 -- =============================================================================
 -- SECTION 4: Company — change id TEXT→INTEGER, add uid PK, add columns
+-- Production has ~4500 rows with CUID text PKs and cw_CompanyId integers
+-- that must be preserved as uid and id respectively.
 -- =============================================================================
 
 -- Drop FKs that reference Company by old id
@@ -229,12 +231,18 @@ DROP INDEX IF EXISTS "Company_cw_CompanyId_key";
 DROP INDEX IF EXISTS "Company_cw_Identifier_key";
 
 DO $$ BEGIN
-  -- Add uid PK column if missing
+  -- Step 1: Add uid as NULLABLE (no default) so existing rows stay NULL temporarily
   IF NOT EXISTS (SELECT 1 FROM information_schema.columns WHERE table_name = 'Company' AND column_name = 'uid') THEN
-    ALTER TABLE "Company" ADD COLUMN "uid" TEXT NOT NULL DEFAULT '';
+    ALTER TABLE "Company" ADD COLUMN "uid" TEXT;
   END IF;
 
-  -- Swap PK from id to uid
+  -- Step 2: Populate uid from the old text PK (old id was a CUID — it becomes uid)
+  UPDATE "Company" SET "uid" = "id" WHERE "uid" IS NULL;
+
+  -- Step 3: Now make uid NOT NULL (all rows are populated)
+  ALTER TABLE "Company" ALTER COLUMN "uid" SET NOT NULL;
+
+  -- Step 4: Swap PK from id (text) to uid (text)
   IF EXISTS (
     SELECT 1 FROM information_schema.table_constraints tc
     JOIN information_schema.key_column_usage kcu ON tc.constraint_name = kcu.constraint_name
@@ -244,7 +252,8 @@ DO $$ BEGIN
     ALTER TABLE "Company" ADD CONSTRAINT "Company_pkey" PRIMARY KEY ("uid");
   END IF;
 
-  -- Change id from TEXT to INTEGER
+  -- Step 5: Change id from TEXT to INTEGER
+  -- NOTE: do this BEFORE dropping cw_CompanyId so we can populate from it below
   IF EXISTS (
     SELECT 1 FROM information_schema.columns
     WHERE table_name = 'Company' AND column_name = 'id' AND data_type = 'text'
@@ -253,7 +262,12 @@ DO $$ BEGIN
     ALTER TABLE "Company" ADD COLUMN "id" INTEGER;
   END IF;
 
-  -- Drop old CW-specific columns
+  -- Step 6: Populate new integer id from cw_CompanyId (CW integer company id)
+  IF EXISTS (SELECT 1 FROM information_schema.columns WHERE table_name = 'Company' AND column_name = 'cw_CompanyId') THEN
+    UPDATE "Company" SET "id" = "cw_CompanyId" WHERE "id" IS NULL;
+  END IF;
+
+  -- Step 7: Drop old CW-specific columns (data now in id and uid)
   IF EXISTS (SELECT 1 FROM information_schema.columns WHERE table_name = 'Company' AND column_name = 'cw_CompanyId') THEN
     ALTER TABLE "Company" DROP COLUMN "cw_CompanyId";
   END IF;
@@ -261,7 +275,7 @@ DO $$ BEGIN
     ALTER TABLE "Company" DROP COLUMN "cw_Identifier";
   END IF;
 
-  -- Add new columns
+  -- Step 8: Add new columns
   IF NOT EXISTS (SELECT 1 FROM information_schema.columns WHERE table_name = 'Company' AND column_name = 'dateDeleted') THEN
     ALTER TABLE "Company" ADD COLUMN "dateDeleted" TIMESTAMP(3);
   END IF;
@@ -291,10 +305,15 @@ DO $$ BEGIN
   END IF;
 END $$;
 
+-- Make Company.id NOT NULL (all rows were populated from cw_CompanyId above)
+ALTER TABLE "Company" ALTER COLUMN "id" SET NOT NULL;
 CREATE UNIQUE INDEX IF NOT EXISTS "Company_id_key" ON "Company"("id");
 
 -- =============================================================================
--- SECTION 5: UnifiSite — change companyId from TEXT to INTEGER
+-- SECTION 5: UnifiSite — change companyId from TEXT to INTEGER (data migration)
+-- Production has ~180 rows where companyId (text) = Company.uid (the old text
+-- PK that was copied into uid in Section 4). We join on Company.uid to get
+-- the new integer Company.id and preserve the relationship.
 -- =============================================================================
 
 DO $$ BEGIN
@@ -302,8 +321,24 @@ DO $$ BEGIN
     SELECT 1 FROM information_schema.columns
     WHERE table_name = 'UnifiSite' AND column_name = 'companyId' AND data_type = 'text'
   ) THEN
+    -- Add a temporary integer column to hold the mapped value
+    IF NOT EXISTS (SELECT 1 FROM information_schema.columns WHERE table_name = 'UnifiSite' AND column_name = 'companyId_int') THEN
+      ALTER TABLE "UnifiSite" ADD COLUMN "companyId_int" INTEGER;
+    END IF;
+    -- Map old text companyId (= Company.uid) → new integer Company.id
+    UPDATE "UnifiSite" us
+    SET "companyId_int" = c."id"
+    FROM "Company" c
+    WHERE c."uid" = us."companyId";
+    -- Replace old text column with the populated integer column
     ALTER TABLE "UnifiSite" DROP COLUMN "companyId";
-    ALTER TABLE "UnifiSite" ADD COLUMN "companyId" INTEGER;
+    ALTER TABLE "UnifiSite" RENAME COLUMN "companyId_int" TO "companyId";
+  ELSIF EXISTS (
+    SELECT 1 FROM information_schema.columns
+    WHERE table_name = 'UnifiSite' AND column_name = 'companyId_int'
+  ) THEN
+    -- Edge case: int column added but not renamed (interrupted previous run)
+    ALTER TABLE "UnifiSite" RENAME COLUMN "companyId_int" TO "companyId";
   END IF;
 END $$;
 
@@ -1467,7 +1502,8 @@ ALTER TABLE "CatalogItem" ALTER COLUMN "uid" DROP DEFAULT;
 ALTER TABLE "CatalogItem" ALTER COLUMN "id" SET NOT NULL;
 ALTER TABLE "CatalogItem" ALTER COLUMN "subcategoryId" DROP DEFAULT;
 
--- Company: drop defaults, enforce NOT NULL
+-- Company: uid was added nullable (no default), id was made NOT NULL in Section 4.
+-- These are no-ops but kept for safety on fresh DBs.
 ALTER TABLE "Company" ALTER COLUMN "uid" DROP DEFAULT;
 ALTER TABLE "Company" ALTER COLUMN "id" SET NOT NULL;
 

api/src/modules/pdf-utils/generateQuote.ts

@@ -1,5 +1,5 @@
 import PdfPrinter from "pdfmake/src/Printer";
-import { readFileSync } from "node:fs";
+import { existsSync, readFileSync } from "node:fs";
 import { join } from "node:path";
 
 export interface QuoteLineItem {
@@ -110,7 +110,26 @@ const COMPANY = {
 
 const DEFAULT_LOGO_PATH = join(process.cwd(), "logo.png");
 
-const fontDir = join(process.cwd(), "node_modules/pdfmake/build/fonts/Roboto");
+function resolveRobotoFontDir(): string {
+  const candidates = [
+    join(process.cwd(), "node_modules/pdfmake/build/fonts/Roboto"),
+    join(import.meta.dir, "../../../node_modules/pdfmake/build/fonts/Roboto"),
+    join("/app/node_modules/pdfmake/build/fonts/Roboto"),
+    join("/app/api/node_modules/pdfmake/build/fonts/Roboto"),
+  ];
+
+  for (const dir of candidates) {
+    if (existsSync(join(dir, "Roboto-Medium.ttf"))) {
+      return dir;
+    }
+  }
+
+  throw new Error(
+    `[pdf] Could not locate pdfmake Roboto fonts. Checked: ${candidates.join(", ")}`
+  );
+}
+
+const fontDir = resolveRobotoFontDir();
 const fonts = {
   Roboto: {
     normal: join(fontDir, "Roboto-Regular.ttf"),

api/src/modules/workers/coms.ts

@@ -1,6 +1,7 @@
 import { Server } from "socket.io";
 import { events, EventTypes } from "../globalEvents";
 import { WorkerQueue } from "./queues";
+import { reserveWorkerId } from "../../workert";
 
 function emitGlobalEvent<K extends keyof EventTypes>(
   name: K,

api/src/modules/workers/dalpuri-sync.ts

@@ -1,5 +1,6 @@
 import { Socket } from "socket.io-client";
 import { executeFullDalpuriSync, executeForcedIncrementalDalpuriSync } from "dalpuri";
+import { prisma } from "../../constants";
 
 /**
  * Execute a full sync from Dalpuri (ConnectWise) to the API database.
@@ -14,5 +15,53 @@ export async function executeFullSync(_workerSocket: Socket): Promise<void> {
  * Called every 5 seconds via PgBoss from the API process interval.
  */
 export async function executeIncrementalSync(): Promise<void> {
-  return executeForcedIncrementalDalpuriSync();
+  let jobRunId: string | undefined;
+
+  try {
+    const run = await prisma.syncJobRun.create({
+      data: {
+        jobType: "INCREMENTAL_SYNC",
+        status: "RUNNING",
+        triggeredBy: "worker",
+        startedAt: new Date(),
+      },
+      select: { id: true },
+    });
+    jobRunId = run.id;
+  } catch (err) {
+    // Sync should still run even if tracking insert fails.
+    console.error("[sync] Failed to create incremental SyncJobRun", err);
+  }
+
+  try {
+    await executeForcedIncrementalDalpuriSync({ jobRunId });
+
+    if (jobRunId) {
+      await prisma.syncJobRun.update({
+        where: { id: jobRunId },
+        data: {
+          status: "COMPLETED",
+          completedAt: new Date(),
+        },
+      });
+    }
+  } catch (err) {
+    if (jobRunId) {
+      const errorSummary = err instanceof Error ? err.message : String(err);
+      await prisma.syncJobRun
+        .update({
+          where: { id: jobRunId },
+          data: {
+            status: "FAILED",
+            completedAt: new Date(),
+            errorSummary: errorSummary.slice(0, 2000),
+          },
+        })
+        .catch(() => {
+          // Best-effort update only.
+        });
+    }
+
+    throw err;
+  }
 }

api/src/modules/workers/incremental-sync.ts

@@ -6,5 +6,17 @@ import { WorkerQueue } from "./queues";
  * Called on an interval from the main API process so it survives worker restarts.
  */
 export async function enqueueIncrementalSync(): Promise<void> {
-  await getBoss().send(WorkerQueue.DALPURI_INCREMENTAL_SYNC, {});
+  const jobId = await getBoss().send(
+    WorkerQueue.DALPURI_INCREMENTAL_SYNC,
+    {
+      enqueuedAt: new Date().toISOString(),
+    },
+    {
+      singletonKey: "dalpuri-incremental-sync",
+    }
+  );
+
+  if (!jobId) {
+    console.debug("[interval] DALPURI_INCREMENTAL_SYNC already pending or active");
+  }
 }

api/src/modules/workers/jobFactory.ts

@@ -32,6 +32,8 @@ export async function createWorkerJob<T>(
   queueType: WorkerQueue,
   workFn: (workerSocket: Socket) => Promise<T>,
 ): Promise<T> {
+  const managerUrl = process.env.MANAGER_SOCKET_URL ?? "http://localhost:8671";
+
   return new Promise((resolve, reject) => {
     // Request a worker ID and namespace from the manager
     socket.emit(
@@ -53,7 +55,7 @@ export async function createWorkerJob<T>(
         }
 
         // Connect to the worker-specific namespace
-        const workerSocket = io(`http://localhost:8671/worker-${workerId}`, {
+        const workerSocket = io(`${managerUrl}/worker-${workerId}`, {
           reconnection: false,
         });
 

api/src/workert.ts

@@ -163,17 +163,45 @@ if (import.meta.main) {
   // Register job handler for DALPURI_FULL_SYNC
   const { enqueueDalpuriFullSync } = await import("./modules/workers/sync-manager");
   const { executeIncrementalSync } = await import("./modules/workers/dalpuri-sync");
+  const { enqueueIncrementalSync } = await import("./modules/workers/incremental-sync");
   await boss.work(WorkerQueue.DALPURI_FULL_SYNC, async () => {
     const socket = await ensureManagerSocketReady();
-    await enqueueDalpuriFullSync();
+    await enqueueDalpuriFullSync(socket);
   });
   console.log("[worker] Registered DALPURI_FULL_SYNC job handler");
 
   await boss.work(WorkerQueue.DALPURI_INCREMENTAL_SYNC, async () => {
-    await executeIncrementalSync();
+    const startedAt = Date.now();
+    console.log("[worker] DALPURI_INCREMENTAL_SYNC started");
+    try {
+      await executeIncrementalSync();
+      console.log(
+        `[worker] DALPURI_INCREMENTAL_SYNC completed in ${Date.now() - startedAt}ms`
+      );
+    } catch (err) {
+      console.error(
+        `[worker] DALPURI_INCREMENTAL_SYNC failed in ${Date.now() - startedAt}ms`,
+        err
+      );
+      throw err;
+    }
   });
   console.log("[worker] Registered DALPURI_INCREMENTAL_SYNC job handler");
 
+  const enqueueIncrementalWithLogging = () => {
+    enqueueIncrementalSync().catch((err) => {
+      console.error(
+        `[worker] interval enqueueIncrementalSync failed: ${err?.message ?? err}`
+      );
+    });
+  };
+
+  // Keep a worker-local 5s scheduler so incremental sync continues even when
+  // API interval scheduling is unavailable.
+  enqueueIncrementalWithLogging();
+  setInterval(enqueueIncrementalWithLogging, 5_000);
+  console.log("[worker] Started 5-second incremental enqueue interval");
+
   // Register job handler for REFRESH_SALES_METRICS
   const { executeSalesMetricsRefresh } = await import("./modules/workers/sales-metrics");
   await boss.work(WorkerQueue.REFRESH_SALES_METRICS, async (jobs) => {

dalpuri/kubernetes/sync-job.yaml

@@ -0,0 +1,33 @@
+apiVersion: batch/v1
+kind: Job
+metadata:
+  name: dalpuri-sync-RELEASE_TAG
+  namespace: optima
+  labels:
+    app: dalpuri-sync
+spec:
+  backoffLimit: 0
+  ttlSecondsAfterFinished: 86400
+  activeDeadlineSeconds: 7200
+  template:
+    metadata:
+      labels:
+        app: dalpuri-sync
+    spec:
+      containers:
+        - name: sync
+          image: ghcr.io/horizonstacksoftware/optima-dalpuri-sync:RELEASE_TAG
+          env:
+            - name: CW_DATABASE_URL
+              valueFrom:
+                secretKeyRef:
+                  name: api-env-secret
+                  key: CW_DATABASE_URL
+            - name: API_DATABASE_URL
+              valueFrom:
+                secretKeyRef:
+                  name: api-env-secret
+                  key: DATABASE_URL
+      restartPolicy: Never
+      imagePullSecrets:
+        - name: github-container-registry

dalpuri/src/sync-by-table.ts

@@ -294,6 +294,22 @@ const refreshContextFromApi = async (
     }
   }
 
+  const cwMembers = await apiPrisma.cwMember.findMany({
+    select: { cwMemberId: true, identifier: true },
+  });
+  for (const member of cwMembers) {
+    if (
+      member.cwMemberId != null &&
+      member.identifier &&
+      !context.userIdentifiersByMemberRecId.has(member.cwMemberId)
+    ) {
+      context.userIdentifiersByMemberRecId.set(
+        member.cwMemberId,
+        member.identifier
+      );
+    }
+  }
+
   for (const board of boards) {
     context.serviceTicketBoardUidsById.set(board.id, board.uid);
   }
@@ -426,6 +442,12 @@ const sanitizeModelData = (
     ) {
       sanitized.statusId = null;
     }
+    if (
+      sanitized.locationId != null &&
+      !context.corporateLocationIds.has(sanitized.locationId as number)
+    ) {
+      sanitized.locationId = null;
+    }
   }
 
   if (targetModel === "schedule") {
@@ -734,6 +756,11 @@ const getConfigForTable = (table: string): SyncTableConfig | null => {
               secondarySalesFlag: true,
             },
           },
+          soOppStatus: {
+            select: {
+              closedFlag: true,
+            },
+          },
         },
       },
     },

dalpuri/src/sync.ts

@@ -75,6 +75,136 @@ type DeleteResult = {
 
 let incrementalDeleteStepIndex = 0;
 
+const CRITICAL_INCREMENTAL_RECONCILE_TABLES = new Set([
+  "Companies",
+  "Company Addresses",
+  "Contacts",
+  "Opportunities",
+]);
+
+const CRITICAL_CW_WATERMARK_TABLES = new Set([
+  "Companies",
+  "Company Addresses",
+  "Contacts",
+  "Opportunities",
+]);
+
+const criticalFullSyncIntervalMinutes = Math.max(
+  1,
+  Number.parseInt(
+    process.env.DALPURI_CRITICAL_FULL_SYNC_INTERVAL_MINUTES ?? "15",
+    10
+  ) || 15
+);
+
+const CRITICAL_FULL_SYNC_INTERVAL_MS =
+  criticalFullSyncIntervalMinutes * 60 * 1000;
+
+const criticalCwWatermarkOverlapSeconds = Math.max(
+  5,
+  Number.parseInt(
+    process.env.DALPURI_CRITICAL_CW_WATERMARK_OVERLAP_SECONDS ?? "60",
+    10
+  ) || 60
+);
+
+const CRITICAL_CW_WATERMARK_OVERLAP_MS =
+  criticalCwWatermarkOverlapSeconds * 1000;
+
+const criticalCwDeltaLimit = Math.max(
+  100,
+  Number.parseInt(process.env.DALPURI_CRITICAL_CW_DELTA_LIMIT ?? "5000", 10) ||
+    5000
+);
+
+const lastCriticalFullSyncByStep = new Map<string, number>();
+const lastCriticalCwWatermarkByStep = new Map<string, Date>();
+
+const shouldForceCriticalFullSync = (
+  step: Step,
+  forceIncremental: boolean
+): boolean => {
+  if (!forceIncremental) return false;
+  if (!CRITICAL_INCREMENTAL_RECONCILE_TABLES.has(step.name)) return false;
+
+  const now = Date.now();
+  const last = lastCriticalFullSyncByStep.get(step.name) ?? 0;
+
+  if (now - last < CRITICAL_FULL_SYNC_INTERVAL_MS) {
+    return false;
+  }
+
+  lastCriticalFullSyncByStep.set(step.name, now);
+  return true;
+};
+
+const computeCriticalCwWatermarkDecision = async (
+  cwPrisma: CwPrismaClient,
+  step: Step,
+  forceIncremental: boolean
+): Promise<SmartSyncDecision | null> => {
+  if (!forceIncremental) return null;
+  if (!CRITICAL_CW_WATERMARK_TABLES.has(step.name)) return null;
+
+  const cwDelegate = (
+    cwPrisma as unknown as Record<string, { findMany: Function } | undefined>
+  )[step.sourceModel];
+
+  if (!cwDelegate) {
+    return null;
+  }
+
+  const existingWhere =
+    (step.sourceArgs as Record<string, unknown> | undefined)?.where ?? {};
+
+  const lastWatermark = lastCriticalCwWatermarkByStep.get(step.name);
+  const lowerBound = lastWatermark
+    ? new Date(lastWatermark.getTime() - CRITICAL_CW_WATERMARK_OVERLAP_MS)
+    : new Date(Date.now() - CRITICAL_CW_WATERMARK_OVERLAP_MS);
+
+  const rows = (await cwDelegate.findMany({
+    select: {
+      [step.sourceIdField]: true,
+      [step.sourceUpdatedField]: true,
+    },
+    where: {
+      ...(existingWhere as Record<string, unknown>),
+      [step.sourceUpdatedField]: {
+        gte: lowerBound,
+      },
+    },
+    orderBy: { [step.sourceUpdatedField]: "asc" },
+    take: criticalCwDeltaLimit,
+  })) as Row[];
+
+  if (rows.length >= criticalCwDeltaLimit) {
+    console.warn(
+      `  [smart-sync][critical-watermark] ${step.name}: delta reached limit (${criticalCwDeltaLimit}), forcing full sync`
+    );
+    return { mode: "full", differences: [] };
+  }
+
+  if (rows.length > 0) {
+    const latest = rows[rows.length - 1][step.sourceUpdatedField] as Date | null;
+    if (latest) {
+      lastCriticalCwWatermarkByStep.set(step.name, latest);
+    }
+  } else if (!lastWatermark) {
+    lastCriticalCwWatermarkByStep.set(step.name, new Date());
+  }
+
+  const sourceIds = rows.map((r) => r[step.sourceIdField] as number);
+  console.log(
+    `  [smart-sync][critical-watermark] ${step.name}: ${sourceIds.length} ids since ${lowerBound.toISOString()}`
+  );
+
+  return {
+    mode: "incremental",
+    sourceIds,
+    differences: [],
+  };
+};
+
 const parseEnvFile = (path: string): Record<string, string> => {
   const envData = readFileSync(path, "utf8");
   const out: Record<string, string> = {};
@@ -107,6 +237,20 @@ const resolveApiDatabaseUrl = (): string => {
   if (process.env.OPTIMA_API_DATABASE_URL)
     return process.env.OPTIMA_API_DATABASE_URL;
 
+  // Worker/runtime fallback:
+  // In Kubernetes we often provide CW via CW_DATABASE_URL and API Postgres via
+  // DATABASE_URL. Only use DATABASE_URL as API when we can safely infer that.
+  if (process.env.CW_DATABASE_URL && process.env.DATABASE_URL) {
+    return process.env.DATABASE_URL;
+  }
+
+  if (
+    process.env.DATABASE_URL &&
+    /^(postgres|postgresql):\/\//i.test(process.env.DATABASE_URL)
+  ) {
+    return process.env.DATABASE_URL;
+  }
+
   const candidates = [
     resolve(import.meta.dir, "../../api/.env"),
     resolve(process.cwd(), "../api/.env"),
@@ -323,6 +467,22 @@ const refreshContextFromApi = async (
     }
   }
 
+  const cwMembers = await apiPrisma.cwMember.findMany({
+    select: { cwMemberId: true, identifier: true },
+  });
+  for (const member of cwMembers) {
+    if (
+      member.cwMemberId != null &&
+      member.identifier &&
+      !context.userIdentifiersByMemberRecId.has(member.cwMemberId)
+    ) {
+      context.userIdentifiersByMemberRecId.set(
+        member.cwMemberId,
+        member.identifier
+      );
+    }
+  }
+
   for (const board of boards) {
     context.serviceTicketBoardUidsById.set(board.id, board.uid);
   }
@@ -636,6 +796,13 @@ const sanitizeModelData = (
     ) {
       sanitized.stageId = null;
     }
+    // Nullify locationId if the corporate location doesn't exist
+    if (
+      sanitized.locationId != null &&
+      !context.corporateLocationIds.has(sanitized.locationId as number)
+    ) {
+      sanitized.locationId = null;
+    }
     // Nullify taxCodeId if the tax code hasn't synced yet
     if (
       sanitized.taxCodeId != null &&
@@ -1328,6 +1495,15 @@ export const executeFullDalpuriSync = async (options?: {
   const isTimedOut = () => Date.now() - syncStartTime > timeoutMs;
 
   const steps: Step[] = [
+    {
+      name: "CW Members",
+      sourceModel: "member",
+      targetModel: "cwMember",
+      translation: cwMemberTranslation as unknown as AnyTranslation,
+      uniqueField: "cwMemberId",
+      sourceIdField: "memberRecId",
+      sourceUpdatedField: "lastUpdatedUtc",
+    },
     {
       name: "Users",
       sourceModel: "member",
@@ -1342,15 +1518,6 @@ export const executeFullDalpuriSync = async (options?: {
         },
       },
     },
-    {
-      name: "CW Members",
-      sourceModel: "member",
-      targetModel: "cwMember",
-      translation: cwMemberTranslation as unknown as AnyTranslation,
-      uniqueField: "cwMemberId",
-      sourceIdField: "memberRecId",
-      sourceUpdatedField: "lastUpdatedUtc",
-    },
     {
       name: "Companies",
       sourceModel: "company",
@@ -1585,6 +1752,11 @@ export const executeFullDalpuriSync = async (options?: {
               secondarySalesFlag: true,
             },
           },
+          soOppStatus: {
+            select: {
+              closedFlag: true,
+            },
+          },
         },
       },
     },
@@ -1729,19 +1901,40 @@ export const executeFullDalpuriSync = async (options?: {
         step,
         forceIncremental
       );
+      const criticalWatermarkDecision = await computeCriticalCwWatermarkDecision(
+        cwPrisma,
+        step,
+        forceIncremental
+      );
+      const forceCriticalFullSync = shouldForceCriticalFullSync(
+        step,
+        forceIncremental
+      );
+      const effectiveDecision = forceCriticalFullSync
+        ? ({ mode: "full", differences: decision.differences } as SmartSyncDecision)
+        : criticalWatermarkDecision ?? decision;
+
+      if (forceCriticalFullSync) {
+        console.log(
+          `  [smart-sync][forced-full] ${step.name}: forcing periodic full reconciliation every ${criticalFullSyncIntervalMinutes}m`
+        );
+      }
+
       const sourceIdsFilter =
-        decision.mode === "incremental" ? decision.sourceIds : undefined;
+        effectiveDecision.mode === "incremental"
+          ? effectiveDecision.sourceIds
+          : undefined;
       console.log(
         `  [smart-sync]${forceIncremental ? "[forced]" : ""} mode=${
-          decision.mode
+          effectiveDecision.mode
         }${
-          decision.mode === "incremental"
-            ? ` (${decision.sourceIds.length} ids)`
+          effectiveDecision.mode === "incremental"
+            ? ` (${effectiveDecision.sourceIds.length} ids)`
             : ""
         }`
       );
       if (logAllDifferences) {
-        logAllSmartSyncDifferences(step, decision.differences);
+        logAllSmartSyncDifferences(step, effectiveDecision.differences);
       }
       const result = await syncStep(
         cwPrisma,
@@ -1763,7 +1956,7 @@ export const executeFullDalpuriSync = async (options?: {
 
       await writeStepLog(
         step.name,
-        decision.mode,
+        effectiveDecision.mode,
         result,
         { deleted: 0, failed: 0 },
         Date.now() - stepStart
@@ -1860,8 +2053,12 @@ export const executeForcedIncrementalDalpuriSync = async (options?: {
 };
 
 if (import.meta.main) {
-  executeFullDalpuriSync().catch((error) => {
-    console.error("CW -> API sync failed:", error);
-    process.exit(1);
-  });
+  executeFullDalpuriSync()
+    .then(() => {
+      process.exit(0);
+    })
+    .catch((error) => {
+      console.error("CW -> API sync failed:", error);
+      process.exit(1);
+    });
 }

dalpuri/src/translations/opportunity.ts

@@ -1,6 +1,7 @@
 import {
   Opportunity as CwOpportunity,
   OpportunityMember as CwOpportunityMember,
+  SoOppStatus as CwSoOppStatus,
 } from "../../generated/prisma/client";
 import { OpportunityInterest } from "../../../api/generated/prisma/client";
 import { Translation, skipRow } from "./types";
@@ -30,6 +31,7 @@ type ApiOpportunityRecord = {
   dateBecameLead?: Date | null;
   closedDate?: Date | null;
   closedFlag: boolean;
+  locationId?: number | null;
   closedById?: string | null;
   updatedBy: string;
   eneteredBy: string;
@@ -42,6 +44,7 @@ type CwOpportunityWithMembers = CwOpportunity & {
     CwOpportunityMember,
     "memberRecId" | "primarySalesFlag" | "secondarySalesFlag"
   >[];
+  soOppStatus?: Pick<CwSoOppStatus, "closedFlag"> | null;
 };
 
 const toInterest = (value: number | null): OpportunityInterest | null => {
@@ -119,13 +122,19 @@ export const opportunityTranslation: Translation<
     },
     { from: "companyRecId", to: "companyId" },
     { from: "contactRecId", to: "contactId" },
+    { from: "ownerLevelRecId", to: "locationId" },
     { from: "companyAddressRecId", to: "siteId" },
     { from: "poNumber", to: "customerPO" },
     { from: "dateCloseExpected", to: "expectedCloseDate" },
     { from: "datePipelineChange", to: "pipelineChangeDate" },
     { from: "dateBecameLead", to: "dateBecameLead" },
     { from: "dateClosed", to: "closedDate" },
-    { from: "oldCloseFlag", to: "closedFlag" },
+    {
+      from: "oldCloseFlag",
+      to: "closedFlag",
+      process: (_value, _context, row) =>
+        row.soOppStatus?.closedFlag ?? row.oldCloseFlag ?? false,
+    },
     { from: "closedBy", to: "closedById" },
     {
       from: "updatedBy",