[Bug] Weekly usage limit reset early, losing 36% of quota before scheduled reset time

Resolved 💬 7 comments Opened Apr 23, 2026 by mandeep1729 Closed Jun 4, 2026

<img width="857" height="532" alt="Image" src="https://github.com/user-attachments/assets/570f24a3-5e2a-4a62-a310-7be65ab21cb8" />

<img width="965" height="667" alt="Image" src="https://github.com/user-attachments/assets/fdcb51b3-2957-4479-8631-583ad7a69953" />

Bug Description
Title: [BUG] Weekly usage limit reset ~1h 40min before scheduled reset time — 36% of weekly quota lost

Preflight Checklist

  • [x] I have searched existing issues (related: #49616, #41055, #51222)
  • [x] This is a single bug report
  • [x] I am using the latest version of Claude Code (v2.1.118)

What's Wrong?

My weekly usage limit reset approximately 1 hour 40 minutes before the scheduled reset time shown in /status → Usage. I lost ~36% of my remaining weekly quota across all models as a result.

The weekly limit should reset at the scheduled time shown in the Usage tab, not before.

Evidence (two screenshots, same day, same machine):

Screenshot 1 — 9:06 AM PT, April 23, 2026:

  • Current week (all models): 64% used
  • Resets: 12pm (America/Los_Angeles)
  • Current session: 5% used

Screenshot 2 — 10:21 AM PT, April 23, 2026 (~1h 40min before scheduled 12pm reset):

  • Current week (all models): 0% used (bar empty)
  • Resets: Apr 28, 1:59pm (America/Los_Angeles) — a brand-new 7-day window had already started
  • Current week (Sonnet only): 0% used
  • Current session: 10% used, resets 2:59pm

The new reset date of April 28, 1:59pm implies the new week began at approximately 1:59pm PT on April 21 (7 days earlier) — but the reset fired and the counter zeroed out at approximately 10:21 AM PT on April 23, while I was still entitled to my remaining 36% of the previous week.

Expected behavior:

The weekly usage limit should reset exactly at the displayed reset time (12pm PT on April 23), not earlier. Remaining quota in the current week should remain available until the actual scheduled reset.

Actual behavior:

The reset fired at approximately 10:21 AM PT, roughly 1 hour 40 minutes early. Remaining 36% of weekly quota was wiped without being usable.

Environment:

  • Claude Code version: 2.1.118
  • Model: Opus 4.7 (1M context)
  • Plan: Claude Max
  • OS: Ubuntu Linux
  • Platform: CLI (terminal)
  • Timezone: America/Los_Angeles

Impact:

~36% of weekly quota across all models lost. Requesting quota credit for the lost allocation and an engineering fix so this does not recur.

Related issues (this is a recurring pattern):

  • #49616 (weekly limit reset 4+ hours early, April 17, 2026)
  • #41055 (weekly reset failures on Max plan)
  • #51222 (reset time display inconsistencies)

I have both screenshots available and will attach them to this issue. Both images are ~/Downloads/bugs

Environment Info

  • Platform: linux
  • Terminal: gnome-terminal
  • Version: 2.1.118
  • Feedback ID: 480cf030-26d1-41dd-ad72-6293c1a38d0f

Errors

[{"error":"Error: NON-FATAL: Lock acquisition failed for /home/mandeep/.local/share/claude/versions/2.1.118 (expected in multi-process scenarios)\n    at ii$ (/$bunfs/root/src/entrypoints/cli.js:2758:2177)\n    at R46 (/$bunfs/root/src/entrypoints/cli.js:2758:1257)\n    at processTicksAndRejections (native:7:39)","timestamp":"2026-04-23T17:14:10.457Z"}]

View original on GitHub ↗

This issue has 7 comments on GitHub. Read the full discussion on GitHub ↗