I am writing to report a critical flaw in your system that is actively draining my premium quotas without providing the service I am paying for. I consistently receive “Our servers are experiencing high traffic right now” errors, which abruptly terminate the agent. The core issue is that every time I am forced to click “Retry”, my Claude Opus and Sonnet quotas are consumed, despite the fact that no actual analysis or processing was completed.
I pay $200 a month for this platform, and your current infrastructure is severely failing to deliver. Instead of accelerating my workflow, it is destroying my coding momentum and morale. Having my limited tokens unjustly depleted because of your server inadequacy is completely unacceptable. It is effectively stripping me of the resources I paid for and represents a severe breach of user trust.
I demand that this issue be resolved immediately so that server-side failures do not deduct from user quotas. If your infrastructure cannot support paying users, you need to openly acknowledge this failure, issue a full refund of my subscription, and apologize for offering an under-resourced, inadequate product. I expect a definitive and immediate resolution, not automated excuses.
Initiate a coding task or prompt the agent using Claude Opus 4.6 or Sonnet 4.6.
The process halts, and the agent returns the error: “Our servers are experiencing high traffic right now”.
Click the “Retry” button to resume the prompt.
Repeat this multiple times as the server continuously fails.
Navigate to the Models tab and observe that the premium model quotas are significantly depleted, even though the agent failed to execute or complete the requested task.
