[Bug] API Connection Failures on Poor Network Conditions

[Makswell77713]

Bug Description Every request I have a problem - Unable to connect to API due to poor internet connection

Environment Info

• Platform: darwin
• Terminal: intellij
• Version: 2.0.50
• Feedback ID: 1d40ff1f-8032-4a71-997c-7477137068e6

Errors

[{"error":"Error: NON-FATAL: Lock acquisition failed for /Users/apple/.local/share/claude/versions/2.0.50 (expected in multi-process scenarios)\n    at JL0 (/$bunfs/root/claude:2641:1622)\n    at g_A (/$bunfs/root/claude:2641:1300)\n    at processTicksAndRejections (native:7:39)","timestamp":"2025-11-24T13:56:56.672Z"},{"error":"Error: Plugin source missing\n    at esB (/$bunfs/root/claude:1470:36819)\n    at esB (/$bunfs/root/claude:1470:37506)\n    at yL8 (/$bunfs/root/claude:1471:4122)\n    at yL8 (/$bunfs/root/claude:1471:4194)\n    at fL8 (/$bunfs/root/claude:1471:1623)\n    at fL8 (/$bunfs/root/claude:1471:1925)\n    at <anonymous> (/$bunfs/root/claude:1472:3002)\n    at <anonymous> (/$bunfs/root/claude:1472:3197)\n    at processTicksAndRejections (native:7:39)","timestamp":"2025-11-24T13:56:57.527Z"},{"error":"Error: Connection error.\n    at makeRequest (/$bunfs/root/claude:1259:4547)\n    at processTicksAndRejections (native:7:39)","timestamp":"2025-11-24T13:56:58.143Z"},{"error":"Error: NON-FATAL: Lock acquisition failed for /Users/apple/.local/share/claude/versions/2.0.50 (expected in multi-process scenarios)\n    at JL0 (/$bunfs/root/claude:2641:1622)\n    at b_A (/$bunfs/root/claude:2639:12709)\n    at async pA7 (/$bunfs/root/claude:2639:13911)\n    at async Ex (/$bunfs/root/claude:2641:236)\n    at async <anonymous> (/$bunfs/root/claude:2641:12469)\n    at processTicksAndRejections (native:7:39)","timestamp":"2025-11-24T13:57:04.958Z"},{"error":"Error: NON-FATAL: Lock acquisition failed for /Users/apple/.local/share/claude/versions/2.0.50 (expected in multi-process scenarios)\n    at JL0 (/$bunfs/root/claude:2641:1622)\n    at b_A (/$bunfs/root/claude:2639:12709)\n    at async pA7 (/$bunfs/root/claude:2639:13911)\n    at async Ex (/$bunfs/root/claude:2641:236)\n    at async <anonymous> (/$bunfs/root/claude:2641:12469)\n    at processTicksAndRejections (native:7:39)","timestamp":"2025-11-24T13:57:13.114Z"}]

[github-actions[bot]]

Found 3 possible duplicate issues:

1. https://github.com/anthropics/claude-code/issues/11924
2. https://github.com/anthropics/claude-code/issues/12177
3. https://github.com/anthropics/claude-code/issues/8030

This issue will be automatically closed as a duplicate in 3 days.

• If your issue is a duplicate, please close it and 👍 the existing issue instead
• To prevent auto-closure, add a comment or 👎 this comment

🤖 Generated with Claude Code

[Makswell77713]

This is not a duplicate. Everything was working perfectly until Saturday evening, but on Sunday and today the same error appears, and it completely blocks the work.

[RainyPixel]

+1

[saeidakbari]

Same issue from today. Unable to connect to API due to poor internet connection.

[freQuensy23-coder]

Same for me. 50 mbit/sec, ping api.anthropic.com works perfectly, starting from end of last week claude cli works very bad

[GabriFedi97]

+1

[freQuensy23-coder]

I analyzed the traffic from Claude Code in more detail and found that the server is not responding to a number of requests to generate a response. I don't understand the reason, but perhaps it's due to an unstable internet connection.

[Makswell77713]

Every request I have a problem - Unable to connect to API due to poor internet connection

[xinghuolk]

I’m experiencing the same issue. Especially when the write tool handles large files or requests that consume a large number of tokens at once, it only shows that a few dozen to a few hundred tokens were used, and then it gets stuck. I tested it: generating a 500-line file works fine, but generating a 1000-line file consistently triggers the bug. After downgrading to 2.0.49, the problem disappears and everything works normally.

[freQuensy23-coder]

https://github.com/anthropics/claude-code/issues/12251#issuecomment-3601829938

Downgrade don't help(

[phmatray]

I’m seeing the same behavior. Curiously, when I share my network connection from my smartphone (mobile hotspot), the error does not occur. The issue only shows up on my regular Wi-Fi connection, even though the connection is stable and other services work normally. This makes me suspect something about routing or network conditions that the new version handles poorly.

[Makswell77713]

🔍 Claude Code Connection Diagnostics

1️⃣ Testing internet connectivity... ✅ Internet connection OK 2️⃣ Testing DNS resolution for api.anthropic.com... ✅ DNS resolution OK IP: 160.79.104.10 3️⃣ Testing HTTPS connection to api.anthropic.com... ✅ HTTPS connection OK 4️⃣ Checking for running Claude processes... ⚠️ Multiple Claude processes detected (8) This may cause 'Lock acquisition failed' errors Running processes:

• PID: 710 | /Applications/Claude.app/Contents/MacOS/Claude
• PID: 1018 | /Applications/Claude.app/Contents/Frameworks/Claude Helper (Renderer).app/Contents/MacOS/Claude
• PID: 2774 | claude
• PID: 1008 | /Applications/Claude.app/Contents/Frameworks/Claude Helper (Renderer).app/Contents/MacOS/Claude
• PID: 1006 | /Applications/Claude.app/Contents/Frameworks/Claude Helper (Renderer).app/Contents/MacOS/Claude
• PID: 998 | /Applications/Claude.app/Contents/Frameworks/Claude Helper.app/Contents/MacOS/Claude Helper
• PID: 996 | /Applications/Claude.app/Contents/Frameworks/Claude Helper (GPU).app/Contents/MacOS/Claude
• PID: 985 | /Applications/Claude.app/Contents/Frameworks/Electron Framework.framework/Helpers/chrome_crashpad_handler --no-rate-limit 5️⃣ Checking system proxy settings... ✅ No proxy configured 6️⃣ Checking macOS firewall status... Firewall is enabled. (State = 1) 7️⃣ Testing Claude API endpoint... ⚠️ Unexpected response: HTTP 405

The problem continues - it appears suddenly and disappears just as suddenly, but it clearly happens regularly. After a full laptop restart it goes away for a short time. But the fact is that I added the program to the firewall exceptions, the ports respond correctly, and the client works if it’s not launched from the command line. Yet something has changed over the last three versions. It feels like the signal is being intentionally blocked until it finally gets through. The challenge is getting through to the support hotline. It doesn’t feel very serious for a paid service. Yes, I have usage limits, but I would like to be able to use them instead of dancing around hoping for a miracle.

[hecoderius]

I have the same issue. I am not sure what is causing it, but it works for me on a different WIFI or on mobile data. On my WIFI I found that I use WPA and WPA2 mixed mode, and I saw on Perplexity that this can be the root cause. I contacted my internet provider and I will try to change this setting. Once I get the change or learn more, I will share an update.