Skip to content

fix(usage_manager): 🐛 prevent stale API responses from resetting quota count (#75) #81

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
b3nw wants to merge 1 commit into
base: dev
Choose a base branch
from
+13 −8
Open

fix(usage_manager): 🐛 prevent stale API responses from resetting quota count (#75) #81

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 13 additions & 8 deletions src/rotator_library/usage_manager.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3244,8 +3244,13 @@ async def update_quota_baseline(
max_requests = model_data.get("quota_max_requests")

# Sync local request count to API's authoritative value
model_data["request_count"] = used_requests
model_data["requests_at_baseline"] = used_requests
# Use max() to prevent API from resetting our count if it returns stale/cached 100%
# The API can only increase our count (if we missed requests), not decrease it
# See: https://github.com/Mirrowel/LLM-API-Key-Proxy/issues/75
current_count = model_data.get("request_count", 0)
synced_count = max(current_count, used_requests)
model_data["request_count"] = synced_count
model_data["requests_at_baseline"] = synced_count

# Update baseline fields
model_data["baseline_remaining_fraction"] = remaining_fraction
Expand All @@ -3254,7 +3259,7 @@ async def update_quota_baseline(
# Update max_requests and quota_display
if max_requests is not None:
model_data["quota_max_requests"] = max_requests
model_data["quota_display"] = f"{used_requests}/{max_requests}"
model_data["quota_display"] = f"{synced_count}/{max_requests}"

# Handle reset_timestamp: only trust it when quota has been used (< 100%)
# API returns garbage reset times for unused quota
Expand Down Expand Up @@ -3339,19 +3344,19 @@ async def update_quota_baseline(
"approx_cost": 0.0,
},
)
# Sync request tracking
other_model_data["request_count"] = used_requests
# Sync request tracking (use synced_count to prevent reset bug)
other_model_data["request_count"] = synced_count
if max_requests is not None:
other_model_data["quota_max_requests"] = max_requests
other_model_data["quota_display"] = (
f"{used_requests}/{max_requests}"
f"{synced_count}/{max_requests}"
)
# Sync baseline fields
other_model_data["baseline_remaining_fraction"] = (
remaining_fraction
)
other_model_data["baseline_fetched_at"] = now_ts
other_model_data["requests_at_baseline"] = used_requests
other_model_data["requests_at_baseline"] = synced_count
# Sync reset timestamp if valid
if valid_reset_ts:
other_model_data["quota_reset_ts"] = reset_timestamp
Expand Down Expand Up @@ -3381,7 +3386,7 @@ async def update_quota_baseline(

lib_logger.debug(
f"Updated quota baseline for {mask_credential(credential)} model={model}: "
f"remaining={remaining_fraction:.2%}, synced_request_count={used_requests}"
f"remaining={remaining_fraction:.2%}, synced_request_count={synced_count}"
)

await self._save_usage()
Expand Down
Loading