[BUG] Recurring OAuth login failures impacting Pro subscribers — request for SLA/compensation policy
Preflight Checklist
- [x] I have searched existing issues and this hasn't been reported yet
- [x] This is a single bug report (please file separate reports for different bugs)
- [x] I am using the latest version of Claude Code
What's Wrong?
<html>
<body>
<!--StartFragment--><html><head></head><body><h1>Recurring OAuth login failures impacting Pro/Max subscribers — need SLA clarity and compensation policy</h1>
<h2>Preflight Checklist</h2>
<ul>
<li>[x] I have searched existing issues and this hasn't been reported yet</li>
<li>[x] This is a single bug report (please file separate reports for different bugs)</li>
<li>[x] I am using the latest version of Claude Code</li>
</ul>
<h2>What's Wrong?</h2>
<p>Claude Code's OAuth login flow has failed repeatedly over the past two weeks, completely blocking Pro/Max subscribers from using the tool they're paying for. This isn't a one-off — it's a recurring pattern that raises questions about service reliability and whether Anthropic has a compensation or SLA policy for paying customers.</p>
<h3>Affected Dates (confirmed via status.claude.com)</h3>
Date | Incident | Duration
-- | -- | --
March 2, 2026 | Elevated errors on claude.ai, console, and Claude Code | ~4 hours
March 11, 2026 | Elevated errors on Claude.ai (including login issues for Claude Code) | Ongoing
<h3>Error Details</h3>
<pre><code>❯ /login
OAuth error: timeout of 15000ms exceeded
</code></pre>
<p>Retrying does not resolve the issue. The OAuth callback to <code>localhost</code> never completes. Browser-side auth is never reached.</p>
<h3>Environment</h3>
<ul>
<li><strong>Claude Code version:</strong> <!-- run claude --version and paste here --></li>
<li><strong>Platform:</strong> Windows 11</li>
<li><strong>Terminal:</strong> PowerShell / Windows Terminal</li>
<li><strong>Auth method:</strong> Claude.ai subscription (Pro)</li>
<li><strong>Network:</strong> No VPN, no firewall blocking localhost</li>
</ul>
<h3>Impact</h3>
<ul>
<li>Complete inability to use Claude Code during outages — OAuth is a single point of failure for subscription-based auth</li>
<li>No fallback mechanism exists for Pro/Max subscribers (API key auth bills against Console credits, a separate payment stream)</li>
<li>Lost productive hours across multiple incidents in a short window</li>
<li>No proactive communication to subscribers about outages or expected resolution times</li>
</ul>
<h3>The Bigger Issue</h3>
<p>This is not just a bug report — it's a reliability and customer trust concern.</p>
<ol>
<li><strong>No SLA or uptime commitment</strong> is published for Pro/Max Claude Code users. Subscribers are paying $20+/month with no guarantees about availability.</li>
<li><strong>No compensation policy exists</strong> (or if it does, it's not documented). Anthropic's Terms of Service state all payments are non-refundable. For a developer tool that people depend on for daily work, this is a gap.</li>
<li><strong>OAuth as a single point of failure</strong> means that when auth servers go down, 100% of subscription-based Claude Code users are locked out simultaneously — unlike API key users who may still have access.</li>
<li><strong>Support responsiveness</strong> for these issues has been poor. Related GitHub issues report receiving only automated replies, with no human follow-up.</li>
</ol>
<h3>What I'd Like to See</h3>
<ul>
<li>[ ] A published SLA or uptime target for Claude Code (even informal)</li>
<li>[ ] A documented compensation policy for extended outages (prorated credits, subscription extensions, etc.)</li>
<li>[ ] A fallback auth mechanism so Pro/Max subscribers aren't completely locked out during OAuth outages (e.g., temporary offline token, cached credentials with grace period)</li>
<li>[ ] Proactive email/notification to paying subscribers when Claude Code auth is down, with an ETA</li>
</ul>
<h3>Related Issues</h3>
<ul>
<li>#20089 — OAuth callback timeout causes "Failed to fetch user roles" error</li>
<li>#33122 — VS Code Extension OAuth Auth Loop on Windows</li>
<li>#29896 — OAuth credentials silently wiped on failed token refresh</li>
<li>#20756 — OAuth Authentication Broken in WSL2</li>
<li>#19078 — OAuth token expired immediately after fresh login</li>
<li>#21601 — Systematic Service Failures — $440 Refund Request</li>
</ul>
<h3>Additional Context</h3>
<p>I rely on Claude Code daily for professional development work. I'm happy to pay for the service — but paying subscribers deserve either reliability guarantees or a clear path to compensation when the service is unavailable for hours at a time, repeatedly.</p></body></html><!--EndFragment-->
</body>
</html>
What Should Happen?
After completing /login, OAuth authentication should succeed reliably within the 15-second timeout window. When Anthropic's auth servers are experiencing issues, Pro/Max subscribers should either have a fallback authentication mechanism or receive proactive notification that the service is degraded, along with an estimated time to resolution. Paying subscribers should not be completely locked out of the tool with no recourse or compensation.
Error Messages/Logs
i see Test2 from cushlabs
⎿ API Error: 401 {"type":"error","error":{"type":"authentication_error","message":"Invalid authentication credentials"},"request_id":"req_011CYwbkzQ1Cm1xsHpiry1ed"} · Please run /login
❯ i see Test2 from cushlabs
⎿ API Error: 401 {"type":"error","error":{"type":"authentication_error","message":"Invalid authentication credentials"},"request_id":"req_011CYwbnjxtvcoTApMHPWwrK"} · Please run /login
❯ okay so this is really weird I just got a ton of text just now it's 10:19 here in Mexico City time on March 11th and I think I just got our test messages confirm what was sent again pls? ⎿ API Error: 401 {"type":"error","error":{"type":"authentication_error","message":"Invalid authentication credentials"},"request_id":"req_011CYwcwqsUC34Nm2KCFbBVe"} · Please run /login
❯ /login
⎿ Login interrupted
❯ /login
────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
Login
OAuth error: Request failed with status code 500
Press Enter to retry.Steps to Reproduce
Here's what to paste:
Open PowerShell or Windows Terminal on Windows 11
Run claude to launch Claude Code
Run /login and select "Claude.ai subscription" auth method
Wait for the OAuth callback — instead of opening a browser, the CLI hangs for 15 seconds
Receive error: OAuth error: timeout of 15000ms exceeded
Press Enter to retry — same result
Repeat 3–4 times — no change
Confirm at status.claude.com that an active incident is in progress affecting Claude Code login
No workaround available for subscription-based auth users
Note: This is not a local environment issue. Network connectivity is confirmed, no VPN or firewall is involved, and the same setup works normally when Anthropic's auth servers are healthy. The issue is 100% correlated with server-side incidents listed on the status page.
Claude Model
Opus
Is this a regression?
Yes, this worked in a previous version
Last Working Version
_No response_
Claude Code Version
claude --version 2.1.72 (Claude Code)
Platform
Anthropic API
Operating System
Windows
Terminal/Shell
Cursor
Additional Information
_No response_
This issue has 9 comments on GitHub. Read the full discussion on GitHub ↗