Feat: rate limiting

[S1ro1]

No description provided.

[Copilot]

Pull request overview

Adds configurable submission rate limiting to the leaderboard system (stored per leaderboard GPU type in Postgres), exposes management/query surfaces via Discord commands and API endpoints, and standardizes several timestamps to UTC.

Changes:

• Add rate_limit_seconds to leaderboard.gpu_type and implement DB helpers to set/get/enforce limits.
• Enforce rate limits during submission preparation and remove the prior hard-coded API-only rate limit.
• Add admin/user-facing commands/endpoints for rate limit configuration + various UTC timestamp updates / formatting changes.

Reviewed changes

Copilot reviewed 13 out of 13 changed files in this pull request and generated 12 comments.

Show a summary per file

File	Description
src/migrations/20260207_01_EW7ve-leaderboard-rate-limits.py	DB migration adding rate_limit_seconds to leaderboard.gpu_type.
src/libkernelbot/utils.py	Makes parsed deadlines timezone-aware (UTC).
src/libkernelbot/submission.py	Calls DB to enforce submission rate limits before accepting a submission.
src/libkernelbot/run_eval.py	Records eval start/end timestamps in UTC.
src/libkernelbot/leaderboard_db.py	Implements rate limit storage and checks in the DB layer.
src/libkernelbot/backend.py	Uses UTC timestamps when creating submissions.
src/kernelbot/cogs/verify_run_cog.py	Updates verify-task leaderboard deadline to UTC (but still calls create_leaderboard incorrectly).
src/kernelbot/cogs/misc_cog.py	Adds a Discord command to view a leaderboard’s rate limits.
src/kernelbot/cogs/admin_cog.py	Adds an admin Discord command to set a leaderboard GPU rate limit.
src/kernelbot/api/main.py	Adds API endpoints to get/set rate limits; adjusts some timestamps/formatting.
src/kernelbot/api/api_utils.py	Removes prior hard-coded rate limit check in request validation.
AGENTS.md	Adds agent config pointer.
.ruff.toml	Increases Ruff line length to 120.

Comments suppressed due to low confidence (1)

src/kernelbot/api/api_utils.py:145

• _run_submission converts all exceptions into HTTP 400 responses. With rate limiting moved into prepare_submission, this will also turn rate-limit errors into 400s. Consider catching KernelBotError explicitly and using its http_code (e.g., 429) when raising HTTPException, while keeping other exceptions as 400/500 as appropriate.

async def _run_submission(submission: SubmissionRequest, mode: SubmissionMode, backend: KernelBackend):
    try:
        req = prepare_submission(submission, backend)
    except Exception as e:
        raise HTTPException(status_code=400, detail=str(e)) from e

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[github-actions]

Coverage report

Click to see where and how coverage changed

File	Statements	Missing	Coverage	Coverage (new stmts)	Lines missing
src/libkernelbot
backend.py
leaderboard_db.py					1134-1137
submission.py
utils.py					64-69
Project Total

This report was generated by python-coverage-comment-action

[msaroufim]

@claude review this code please

[msaroufim]

We probably also want user based rate limits but LGTM