improvement(db): contract workspace file sizes

Author: waleedlatif1

.github/workflows/migrations.yml

@@ -80,7 +80,6 @@ jobs:
               echo "ERROR: db:push needs an interactive rename decision; land it as a versioned migration instead of relying on push." >&2
               exit 1
             fi
-            bun run ./scripts/apply-dev-workspace-file-size-cutover.ts
           else
             echo "Applying versioned migrations (db:migrate)"
             bun run ./scripts/migrate.ts

apps/sim/background/cleanup-soft-deletes.test.ts

@@ -198,26 +198,6 @@ describe('cleanup soft deletes', () => {
     expect(mockDecrementStorageUsageForBillingContextInTx).not.toHaveBeenCalled()
   })
 
-  it('fails before deleting storage when canonical size metadata is missing', async () => {
-    mockSelectRowsByIdChunks
-      .mockResolvedValueOnce([])
-      .mockResolvedValueOnce([])
-      .mockResolvedValueOnce([
-        {
-          id: 'file-missing-size',
-          key: 'workspace/ws-1/file-missing-size',
-          workspaceId: 'ws-1',
-          context: 'workspace',
-          sizeBytes: null,
-        },
-      ])
-
-    await expect(runCleanupSoftDeletes(basePayload)).rejects.toThrow(
-      'Workspace file is missing canonical size_bytes metadata'
-    )
-    expect(mockDeleteFiles).not.toHaveBeenCalled()
-  })
-
   it('hard-deletes retained documents before deleting an expired knowledge base', async () => {
     mockChunkedBatchDelete.mockImplementationOnce(
       async (options: {

apps/sim/ee/workspace-forking/lib/copy/storage-quota.test.ts

@@ -49,7 +49,7 @@ import {
 } from '@/ee/workspace-forking/lib/copy/storage-quota'
 import { ForkError } from '@/ee/workspace-forking/lib/lineage/authz'
 
-function makeExecutor(total: number | string | null) {
+function makeExecutor(total: number | string) {
   const execute = vi.fn((_query: unknown) => Promise.resolve([{ total }]))
   return { executor: { execute } as unknown as DbOrTx, execute }
 }
@@ -71,8 +71,7 @@ describe('sumForkCopyBytes', () => {
     const compiled = outerQuery.toSQL()
     expect(compiled.sql).toBe('SELECT (? + ?)::bigint AS total')
     const [fileBytes, kbBytes] = compiled.params
-    expect(fileBytes.toSQL().sql).toContain('count(*) FILTER')
-    expect(fileBytes.toSQL().sql).toContain('IS NULL')
+    expect(fileBytes.toSQL().sql).toContain('coalesce(sum')
     expect(fileBytes.toSQL().params).toContainEqual({
       type: 'and',
       conditions: [
@@ -103,17 +102,6 @@ describe('sumForkCopyBytes', () => {
     expect(bytes).toBe(1024)
   })
 
-  it('fails closed when a selected workspace file lacks canonical size metadata', async () => {
-    const { executor } = makeExecutor(null)
-
-    await expect(
-      sumForkCopyBytes(executor, 'src-ws', { fileIds: ['wf-missing-size'] })
-    ).rejects.toMatchObject({
-      message: 'Storage calculation is temporarily unavailable',
-      statusCode: 503,
-    })
-  })
-
   it('runs no query for an empty selection', async () => {
     const { executor, execute } = makeExecutor(0)
 

apps/sim/ee/workspace-forking/lib/copy/storage-quota.ts

@@ -47,11 +47,8 @@ export async function sumForkCopyBytes(
   const fileBytes =
     fileSelectors.length === 0
       ? sql<number>`0`
-      : sql<number | null>`(
-          SELECT CASE
-            WHEN count(*) FILTER (WHERE ${workspaceFiles.sizeBytes} IS NULL) > 0 THEN NULL
-            ELSE coalesce(sum(${workspaceFiles.sizeBytes}), 0)
-          END
+      : sql<number>`(
+          SELECT coalesce(sum(${workspaceFiles.sizeBytes}), 0)
           FROM ${workspaceFiles}
           WHERE ${and(
             fileSelectors.length === 1 ? fileSelectors[0] : or(...fileSelectors),
@@ -77,10 +74,10 @@ export async function sumForkCopyBytes(
             isNotNull(document.storageKey)
           )}
         )`
-  const [row] = await executor.execute<{ total: number | string | null }>(
+  const [row] = await executor.execute<{ total: number | string }>(
     sql`SELECT (${fileBytes} + ${kbBytes})::bigint AS total`
   )
-  if (row?.total == null) {
+  if (!row) {
     throw new ForkError('Storage calculation is temporarily unavailable', 503)
   }
   return Number(row.total)

apps/sim/lib/billing/storage/payer-transfer.test.ts

@@ -102,7 +102,6 @@ interface FakeTransferState {
   users?: Record<string, number>
   workspace: FakeWorkspace
   workspaceFileBytes?: number
-  workspaceFileMissingSizeCount?: number
 }
 
 function createFakeTx(state: FakeTransferState) {
@@ -166,7 +165,6 @@ function createFakeTx(state: FakeTransferState) {
       {
         document_bytes: state.documentBytes ?? 0,
         workspace_file_bytes: state.workspaceFileBytes ?? 0,
-        workspace_file_missing_size_count: state.workspaceFileMissingSizeCount ?? 0,
       },
     ]
   })
@@ -188,7 +186,6 @@ function updateFor(
 
 interface FakeBatchTransferState {
   exactBytes: Record<string, number>
-  missingSizeCounts?: Record<string, number>
   organizations?: Record<string, number>
   users?: Record<string, number>
   workspaces: FakeWorkspace[]
@@ -217,7 +214,6 @@ function createFakeBatchTx(state: FakeBatchTransferState) {
       workspace_id: workspaceId,
       document_bytes: 0,
       workspace_file_bytes: bytes,
-      workspace_file_missing_size_count: state.missingSizeCounts?.[workspaceId] ?? 0,
     }))
   )
 
@@ -535,31 +531,7 @@ describe('changeWorkspaceStoragePayerInTx', () => {
     expect(query.values).not.toContain('workspaceFiles.deletedAt')
     expect(query.values).toContain('document.connectorId')
     expect(query.values).toContain('document.deletedAt')
-    expect(query.values.filter((value) => value === 'workspace-1')).toHaveLength(3)
-  })
-
-  it('fails closed when a billable file is missing canonical size metadata', async () => {
-    const fake = createFakeTx({
-      workspace: {
-        id: 'workspace-1',
-        billedAccountUserId: 'user-1',
-        organizationId: null,
-        storageUsedBytes: 10,
-      },
-      workspaceFileBytes: 10,
-      workspaceFileMissingSizeCount: 1,
-      users: { 'user-1': 10, 'user-2': 0 },
-    })
-
-    await expect(
-      changeWorkspaceStoragePayerInTx(fake.tx, {
-        workspaceId: 'workspace-1',
-        organizationId: null,
-        billedAccountUserId: 'user-2',
-      })
-    ).rejects.toThrow('Workspace workspace-1 has files missing canonical size_bytes metadata')
-
-    expect(fake.updates).toEqual([])
+    expect(query.values.filter((value) => value === 'workspace-1')).toHaveLength(2)
   })
 })
 
@@ -712,34 +684,6 @@ describe('changeWorkspaceStoragePayersInTx', () => {
     expect(fake.updates).toEqual([])
     expect(fake.locks).toEqual([{ ids: ['workspace-a'], table: 'workspace' }])
   })
-
-  it('fails the batch before payer writes when canonical size metadata is missing', async () => {
-    const fake = createFakeBatchTx({
-      exactBytes: { 'workspace-a': 10 },
-      missingSizeCounts: { 'workspace-a': 1 },
-      users: { current: 10, destination: 0 },
-      workspaces: [
-        {
-          id: 'workspace-a',
-          billedAccountUserId: 'current',
-          organizationId: null,
-          storageUsedBytes: 10,
-        },
-      ],
-    })
-
-    await expect(
-      changeWorkspaceStoragePayersInTx(fake.tx, [
-        {
-          workspaceId: 'workspace-a',
-          organizationId: null,
-          billedAccountUserId: 'destination',
-        },
-      ])
-    ).rejects.toThrow('Workspace workspace-a has files missing canonical size_bytes metadata')
-
-    expect(fake.updates).toEqual([])
-  })
 })
 
 describe('changeOrganizationWorkspaceBilledAccountsInTx', () => {

apps/sim/lib/billing/storage/payer-transfer.ts

@@ -17,7 +17,6 @@ interface ExactWorkspaceStorageRow {
   [key: string]: unknown
   document_bytes: number | string
   workspace_file_bytes: number | string
-  workspace_file_missing_size_count: number | string
 }
 
 interface BatchExactWorkspaceStorageRow extends ExactWorkspaceStorageRow {
@@ -88,13 +87,6 @@ async function getExactWorkspaceStorageBytes(tx: DbOrTx, workspaceId: string): P
         WHERE ${workspaceFiles.workspaceId} = ${workspaceId}
           AND ${workspaceFiles.context} = 'workspace'
       ), 0)::bigint AS workspace_file_bytes,
-      (
-        SELECT COUNT(*)
-        FROM ${workspaceFiles}
-        WHERE ${workspaceFiles.workspaceId} = ${workspaceId}
-          AND ${workspaceFiles.context} = 'workspace'
-          AND ${workspaceFiles.sizeBytes} IS NULL
-      )::bigint AS workspace_file_missing_size_count,
       COALESCE((
         SELECT SUM(${document.fileSize}::bigint)
         FROM ${document}
@@ -109,10 +101,6 @@ async function getExactWorkspaceStorageBytes(tx: DbOrTx, workspaceId: string): P
   if (!row) {
     throw new Error(`Could not recompute storage for workspace ${workspaceId}`)
   }
-  if (parseExactBytes(row.workspace_file_missing_size_count, 'missing workspace file size') > 0) {
-    throw new Error(`Workspace ${workspaceId} has files missing canonical size_bytes metadata`)
-  }
-
   const workspaceFileBytes = parseExactBytes(row.workspace_file_bytes, 'workspace file')
   const documentBytes = parseExactBytes(row.document_bytes, 'knowledge document')
   const total = workspaceFileBytes + documentBytes
@@ -178,16 +166,12 @@ async function getExactWorkspaceStorageBytesBatch(
       COALESCE(SUM(storage_by_workspace.workspace_file_bytes), 0)::bigint
         AS workspace_file_bytes,
       COALESCE(SUM(storage_by_workspace.document_bytes), 0)::bigint
-        AS document_bytes,
-      COALESCE(SUM(storage_by_workspace.workspace_file_missing_size_count), 0)::bigint
-        AS workspace_file_missing_size_count
+        AS document_bytes
     FROM (
       SELECT
         ${workspaceFiles.workspaceId} AS workspace_id,
         SUM(${workspaceFiles.sizeBytes}) AS workspace_file_bytes,
-        0::bigint AS document_bytes,
-        COUNT(*) FILTER (WHERE ${workspaceFiles.sizeBytes} IS NULL)::bigint
-          AS workspace_file_missing_size_count
+        0::bigint AS document_bytes
       FROM ${workspaceFiles}
       WHERE ${inArray(workspaceFiles.workspaceId, workspaceIds)}
         AND ${workspaceFiles.context} = 'workspace'
@@ -198,8 +182,7 @@ async function getExactWorkspaceStorageBytesBatch(
       SELECT
         ${knowledgeBase.workspaceId} AS workspace_id,
         0::bigint AS workspace_file_bytes,
-        SUM(${document.fileSize}::bigint) AS document_bytes,
-        0::bigint AS workspace_file_missing_size_count
+        SUM(${document.fileSize}::bigint) AS document_bytes
       FROM ${document}
       INNER JOIN ${knowledgeBase}
         ON ${knowledgeBase.id} = ${document.knowledgeBaseId}
@@ -213,11 +196,6 @@ async function getExactWorkspaceStorageBytesBatch(
   `)
 
   for (const row of rows) {
-    if (parseExactBytes(row.workspace_file_missing_size_count, 'missing workspace file size') > 0) {
-      throw new Error(
-        `Workspace ${row.workspace_id} has files missing canonical size_bytes metadata`
-      )
-    }
     const workspaceFileBytes = parseExactBytes(row.workspace_file_bytes, 'workspace file')
     const documentBytes = parseExactBytes(row.document_bytes, 'knowledge document')
     const total = workspaceFileBytes + documentBytes

apps/sim/lib/uploads/shared/types.ts

@@ -5,17 +5,8 @@
  */
 export const MAX_WORKSPACE_FILE_SIZE = 5 * 1024 * 1024 * 1024
 
-/**
- * Returns the canonical workspace-file byte size after the `size_bytes` cutover.
- *
- * The migration backfills every existing row before the new application image is
- * promoted, and its compatibility trigger fills the column for writes from an old
- * image during rollout. A null therefore indicates migration drift, not a legacy row.
- */
-export function getWorkspaceFileSize(file: { sizeBytes: number | null }): number {
-  if (file.sizeBytes === null) {
-    throw new Error('Workspace file is missing canonical size_bytes metadata')
-  }
+/** Returns a validated workspace-file byte size. */
+export function getWorkspaceFileSize(file: { sizeBytes: number }): number {
   if (!Number.isSafeInteger(file.sizeBytes) || file.sizeBytes < 0) {
     throw new Error(`Invalid workspace file size: ${file.sizeBytes}`)
   }

packages/db/migrations/0309_contract_workspace_file_size.sql

@@ -0,0 +1,14 @@
+ALTER TABLE "workspace_files"
+	ADD CONSTRAINT "workspace_files_size_bytes_not_null_check"
+	CHECK ("size_bytes" IS NOT NULL) NOT VALID;--> statement-breakpoint
+ALTER TABLE "workspace_files"
+	VALIDATE CONSTRAINT "workspace_files_size_bytes_not_null_check";--> statement-breakpoint
+-- migration-safe: contract of #7112 and #7123 — application reads and writes use size_bytes, the backfill is complete, and this PR must merge only after the compatibility release fully drains
+ALTER TABLE "workspace_files" ALTER COLUMN "size_bytes" SET NOT NULL;--> statement-breakpoint
+-- migration-safe: removes the temporary proof constraint created and validated above after PostgreSQL records the equivalent column-level NOT NULL invariant
+ALTER TABLE "workspace_files"
+	DROP CONSTRAINT "workspace_files_size_bytes_not_null_check";--> statement-breakpoint
+DROP TRIGGER IF EXISTS "workspace_files_sync_size_columns" ON "workspace_files";--> statement-breakpoint
+DROP FUNCTION IF EXISTS "sync_workspace_file_size_columns"();--> statement-breakpoint
+-- migration-safe: contract of #7112 and #7123 — no deployed application reader or writer depends on size, and this PR must merge only after the compatibility release fully drains
+ALTER TABLE "workspace_files" DROP COLUMN "size";