D1 database is visible in dashboard and d1 list but Wrangler returns errors 7404 and 10181

[jimviva-bo]

What versions & operating system are you using?

Reported by: Jim Acquaviva

Operating system: Windows 11
Node.js: v24.19.0
npm: 11.17.0
Wrangler tested: 4.121.0 and 4.120.1

The same problem occurs with both Wrangler versions.

Please provide a link to a minimal reproduction

#10897

Describe the Bug

I have an existing Cloudflare D1 database named bcb-inventory. It is visible in the Cloudflare dashboard, appears when I run npx wrangler d1 list, and works correctly through the D1 dashboard Console.

The database is also successfully bound to my currently active Worker as DB, and the live Worker continues to operate.

However, Wrangler cannot access this same database for migrations or validate it during a new deployment.

Running:

npx wrangler d1 migrations list bcb-inventory --remote

returns error 7404 stating that the database cannot be found.

Running:

npm run deploy

uploads the Worker assets successfully but fails while creating the new Worker version. It returns error 10181 stating that the D1 database UUID could not be found.

I confirmed all of the following:

• Wrangler is logged into the correct Cloudflare account.
• The OAuth login was completely refreshed with wrangler logout and wrangler login.
• The database UUID in wrangler.jsonc matches the UUID returned by wrangler d1 list.
• The binding name is DB.
• The database has no special jurisdiction; JSON output reports "jurisdiction": null.
• The problem occurs with Wrangler 4.121.0 and 4.120.1.
• The database remains accessible through the Cloudflare dashboard Console.
• The existing production Worker already has a functioning DB → bcb-inventory binding.
• The failed deployment does not create or activate a new Worker version.

This appears to be an inconsistency between the D1 dashboard/control plane and the API used by Wrangler.

Please provide any relevant error logs

No response

[workers-devprod]

Note

This is an automated, advisory triage report generated by workers-devprod. It is not an official maintainer response — a maintainer will follow up.

🤖 Automated triage report

Issue Triage: #15144

Summary

An existing D1 database (bcb-inventory) that works via the dashboard, wrangler d1 list, and a live Worker binding cannot be found by Wrangler during d1 migrations list --remote (error 7404) or during wrangler deploy (error 10181).

Classification

• Type: Bug
• Component: packages/wrangler (D1 code lives in wrangler)
• Severity: High
• Has Reproduction: No (the "reproduction link" field points to another issue, D1 binding fails with error 10021 for existing database #10897, not a minimal repro repo)
• Quality: Partial

Findings

• Created: not determinable from the truncated context.json (single-line body exceeds the 2000-char read limit; metadata fields such as labels, comments, createdAt, and state were not readable with the available tools)
• Author: jimviva-bo (reporting on behalf of Jim Acquaviva)
• Version: Wrangler 4.121.0 and 4.120.1 (both recent); Node v24.19.0; npm 11.17.0; Windows 11
• Labels: not readable (see note above)
• Comments: not readable (see note above)

Analysis

• Symptom: The same D1 database is reachable through the dashboard Console, wrangler d1 list, and an already-deployed Worker's DB binding, but Wrangler's migration and deploy code paths report the database as not found (D1 API error 7404, deploy error 10181 "D1 database UUID could not be found"). This inconsistency between read paths points to Wrangler/D1 tooling hitting a different account, API surface, or endpoint than the dashboard/d1 list.
• Diagnostics already ruled out by reporter: correct account, fresh logout/login, UUID in wrangler.jsonc matches d1 list, binding name DB, jurisdiction: null, reproduced on two Wrangler versions. This is a thorough report that eliminates the common misconfiguration causes.
• Reproduction gap: The reproduction-link field contains https://github.com/cloudflare/workers-sdk/issues/10897 — a link to another workers-sdk issue rather than a minimal, clonable reproduction. D1 binding fails with error 10021 for existing database #10897 appears to be a prior report in the same "D1 database cannot be found" family, so this may be a duplicate or closely related.
• Component confidence: D1 command handling and deploy-time binding validation both live in packages/wrangler; this is a wrangler/D1 tooling issue, not a workerd runtime issue.
• Not stale / not spam / not user error: Filed against current Wrangler versions with detailed, credible diagnostics that eliminate obvious user error.

Recommendation

Status: NEEDS VERIFICATION

Reasoning: This is a credible, well-diagnosed bug against current Wrangler, but it lacks a true minimal reproduction (the repro field links to issue #10897 instead of a repo) and the account/database context can't be verified from the issue alone. A maintainer should first check whether it duplicates #10897 and then confirm the account/endpoint mismatch.

Action: A maintainer should (1) compare against #10897 and consolidate if it is the same underlying problem; (2) request the failing command's full --log-level debug output including the D1 API request URL and the resolved account_id, to confirm whether Wrangler is querying a different account/endpoint than the dashboard; (3) confirm whether the database was created in or moved between accounts, or has a non-default location/jurisdiction that the deploy path doesn't resolve.

Suggested Labels: (none applied — could not confirm which labels exist in the repo from the available data; a maintainer may add product:d1 and, if a repro is still missing after follow-up, awaiting-response:reporter)

Suggested Comment

Thanks for the detailed report — the diagnostics you've already run are really helpful.

To dig into why the D1 API reports 7404/10181 while the dashboard and d1 list succeed, could you share:

• The full output of npx wrangler d1 migrations list bcb-inventory --remote run with --log-level debug (this will include the exact D1 API URL and resolved account_id Wrangler is calling — please redact any tokens).
• The account_id shown in your wrangler.jsonc (or confirmation of which account is selected) versus the account that owns the database in the dashboard.
• Whether the database was ever created in, or moved between, different Cloudflare accounts, and its location/region.

Also, your reproduction link points to #10897 rather than a minimal repro — is this the same underlying problem as that issue? If so, we can consolidate. If not, a small repo we can clone would help us investigate further.

Structured summary (JSON)

{
  "issueNumber": 15144,
  "title": "D1 database found by dashboard and d1 list but not by wrangler migrations/deploy (error 7404 / 10181)",
  "githubUrl": "https://github.com/cloudflare/workers-sdk/issues/15144",
  "recommendation": "NEEDS VERIFICATION",
  "difficulty": "medium",
  "reasoning": "Credible, well-diagnosed D1 bug against current Wrangler where the database is reachable via dashboard and d1 list but not via migrations/deploy, but it lacks a true minimal reproduction and the account/endpoint context can't be verified from the issue alone.",
  "suggestedAction": "Check whether it duplicates #10897, then request debug-level output showing the D1 API URL and resolved account_id to confirm an account/endpoint mismatch.",
  "hasSuggestedComment": true,
  "suggestedLabels": []
}

[workers-devprod]

• Reproduction: ❌ Could not reproduce. The issue requires a live Cloudflare account with a specific D1 backend inconsistency that cannot be replicated in a sandbox. I verified the Wrangler code paths by running the existing D1 test suite (all passing) and confirmed that error 7404 is propagated from the Cloudflare API query endpoint, while error 10181 does not appear anywhere in the Wrangler codebase, indicating it is a server-side Workers API validation error.