improvement(db): contract workspace file sizes

[waleedlatif1]

Summary

• make workspace_files.size_bytes non-null and remove the legacy size bridge, trigger, and function
• remove the temporary backfill runner and obsolete null-handling paths
• keep this draft unmerged until the compatibility release has fully drained

Type of Change

• Maintenance

Testing

• bun run lint
• bun run check:audits
• bun run type-check
• bun run check:migrations origin/staging
• focused Vitest suites (54 tests)
• bunx drizzle-kit check

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
docs	Ready	Preview	Aug 26, 2026 9:33pm

[waleedlatif1]

@greptile

[waleedlatif1]

@cubic

[cubic-dev-ai]

@cubic

@waleedlatif1 I have started the AI code review. It will take a few minutes to complete.

[greptile-apps]

Greptile Summary

This PR completes the workspace-file-size contract after the compatibility rollout.

• Makes workspace_files.size_bytes non-null and removes the legacy size column, synchronization trigger, and function.
• Removes the completed size backfill migration and development cutover runner.
• Simplifies storage accounting, cleanup, upload, and fork-quota paths to rely on the database-enforced invariant.

Confidence Score: 5/5

The PR appears safe to merge once its explicitly stated compatibility-release drain condition has been satisfied.

The migration validates existing data before enforcing NOT NULL, then removes the compatibility bridge, while changed runtime paths consistently adopt the resulting database contract; no independent blocking or non-blocking defect remains.

Important Files Changed

Filename	Overview
packages/db/migrations/0309_contract_workspace_file_size.sql	Contracts the compatibility schema by validating and enforcing non-null byte sizes before removing the legacy bridge.
packages/db/schema.ts	Updates the Drizzle contract to expose only the canonical non-null sizeBytes column.
packages/db/script-migrations/index.ts	Removes the completed workspace-file-size backfill from the script-migration registry.
.github/workflows/migrations.yml	Removes the temporary development cutover runner after the compatibility phase.
apps/sim/ee/workspace-forking/lib/copy/storage-quota.ts	Simplifies fork-copy byte aggregation to rely on the new database non-null invariant.
apps/sim/lib/billing/storage/payer-transfer.ts	Removes obsolete missing-size checks from exact storage recomputation while retaining aggregate validation.
apps/sim/lib/uploads/shared/types.ts	Tightens the workspace-file-size helper input type to the canonical non-null schema.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart LR
  A[Compatibility release] --> B[Backfill size_bytes]
  B --> C[Drain legacy writers]
  C --> D[Validate size_bytes is non-null]
  D --> E[Set column NOT NULL]
  E --> F[Drop compatibility trigger and function]
  F --> G[Drop legacy size column]
  G --> H[Application paths rely on size_bytes]

Loading

_{Reviews (1): Last reviewed commit: "improvement(db): contract workspace file..." | Re-trigger Greptile}

[cubic-dev-ai]

2 issues found across 19 files

Confidence score: 2/5

• apps/sim/ee/workspace-forking/lib/copy/storage-quota.ts can undercount copied bytes when selected workspace_files rows have size_bytes = NULL, allowing quota admission to exceed the intended limit — handle NULL sizes before summing or admitting the copy.
• packages/db/migrations/0309_contract_workspace_file_size.sql can abort on existing databases containing legacy rows with NULL size_bytes, blocking migration — add an idempotent backfill from the still-present non-null size column before validation.

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="apps/sim/ee/workspace-forking/lib/copy/storage-quota.ts">

<violation number="1" location="apps/sim/ee/workspace-forking/lib/copy/storage-quota.ts:51">
P1: When a selected `workspace_files` row still has `size_bytes = NULL`, PostgreSQL ignores it in `sum`, and this `coalesce` turns an all-NULL selection into zero, allowing quota admission to undercount copied bytes. Preserve the NULL-detection aggregate and make the 503 path handle a null total.

(Based on your team's feedback about failing closed on missing `size_bytes`.) [3f8e6e9d-39ca-4a4a-b8b4-0e6be9b461e0].</violation>
</file>

<file name="packages/db/migrations/0309_contract_workspace_file_size.sql">

<violation number="1" location="packages/db/migrations/0309_contract_workspace_file_size.sql:5">
P1: When an existing database has rows from before `size_bytes` was added, this validation aborts because the migration does not backfill NULLs. Add an idempotent update from the still-present non-null `size` column before validation so self-hosted upgrades do not depend on the removed runner.

(Based on your team's feedback about deployable schema/data migrations.)</violation>
</file>

<sub>Tip: cubic can generate docs of your entire codebase and keep them up to date. Try it here.

Re-trigger cubic</sub>

[cubic-dev-ai]

P1: When a selected workspace_files row still has size_bytes = NULL, PostgreSQL ignores it in sum, and this coalesce turns an all-NULL selection into zero, allowing quota admission to undercount copied bytes. Preserve the NULL-detection aggregate and make the 503 path handle a null total.

(Based on your team's feedback about failing closed on missing size_bytes.) [3f8e6e9d-39ca-4a4a-b8b4-0e6be9b461e0].

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At apps/sim/ee/workspace-forking/lib/copy/storage-quota.ts, line 51:

<comment>When a selected `workspace_files` row still has `size_bytes = NULL`, PostgreSQL ignores it in `sum`, and this `coalesce` turns an all-NULL selection into zero, allowing quota admission to undercount copied bytes. Preserve the NULL-detection aggregate and make the 503 path handle a null total.

(Based on your team's feedback about failing closed on missing `size_bytes`.) [3f8e6e9d-39ca-4a4a-b8b4-0e6be9b461e0].</comment>

<file context>
@@ -47,11 +47,8 @@ export async function sumForkCopyBytes(
-            ELSE coalesce(sum(${workspaceFiles.sizeBytes}), 0)
-          END
+      : sql<number>`(
+          SELECT coalesce(sum(${workspaceFiles.sizeBytes}), 0)
           FROM ${workspaceFiles}
           WHERE ${and(
</file context>

[cubic-dev-ai]

P1: When an existing database has rows from before size_bytes was added, this validation aborts because the migration does not backfill NULLs. Add an idempotent update from the still-present non-null size column before validation so self-hosted upgrades do not depend on the removed runner.

(Based on your team's feedback about deployable schema/data migrations.)

View Feedback

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At packages/db/migrations/0309_contract_workspace_file_size.sql, line 5:

<comment>When an existing database has rows from before `size_bytes` was added, this validation aborts because the migration does not backfill NULLs. Add an idempotent update from the still-present non-null `size` column before validation so self-hosted upgrades do not depend on the removed runner.

(Based on your team's feedback about deployable schema/data migrations.) </comment>

<file context>
@@ -0,0 +1,14 @@
+	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
</file context>

Suggested change

	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
	UPDATE "workspace_files"
	SET "size_bytes" = "size"
	WHERE "size_bytes" IS NULL;--> statement-breakpoint
	ALTER TABLE "workspace_files"
	VALIDATE CONSTRAINT "workspace_files_size_bytes_not_null_check";--> statement-breakpoint