GRatekeeper

GRatekeeper (“GitHub Ratekeeper”) is a lightweight companion for monitoring and controlling GitHub API usage. It gives you:

• A lean, configurable dashboard (gk-dash / gratekeeper-dashboard) that shows every quota bucket, Actions activity, and billing minutes, with quick toggles and themes.
• A drop-in requests.Session wrapper (RateLimitedGitHubClient) that enforces a soft floor before you hit zero, plus optional killswitches for runaway jobs.

Why it’s different from simple retries/backoff:

• Adaptive throttling before zero: While other libraries offer automatic retries and backoff, Gratekeeper slows down your requests before you hit the limit, making it safe to aggressively script your application's API calls without worrying about crossing over into the dreaded secondary rate limit, possibly encurring penalties or prolonged cooldown periods.
• Multi-process friendly: The dashboard/socket lets multiple scripts share live bucket state so one noisy job doesn’t blind the others.
• Last-resort brakes: A killswitch can stop all new requests during incidents, avoiding secondary rate limits.

Quick start

Install (Python 3.10+):

pip install gratekeeper

Use the client:

from gratekeeper import RateLimitedGitHubClient

client = RateLimitedGitHubClient()  # picks up $GITHUB_TOKEN
me = client.get_json("/user")
print(me["login"])

Run the dashboard:

gk-dash --refresh 10 --fetch 30 --actions my-org/my-repo
# prefer a simpler table? add --ui table

Common tasks:

• Customize HTTP: pass token, base_url, user_agent, or bucket="search" to isolate quotas.
• Disable throttling temporarily: enable_ratekeeping=False on the client.
• Themes/presets/bucket toggles/Actions on/off are built into the Textual UI; see the dashboard guide below.

Security and compatibility

• The socket bridge is local-only but unauthenticated; change the path or disable with --socket none / GRATEKEEPER_SOCKET=none on shared hosts. On Windows, the bridge is disabled automatically.
• Works with REST and GraphQL; buckets stay separate unless you intentionally merge them via bucket=....

Where to go next

• Rate-limit math, soft floor, killswitch details: EXPLAINER.md
• Advanced client features (polling, listeners, socket bridge, logging, dev/demos): ADVANCED_USAGE.md
• Dashboard controls (themes, Active-only/manual presets, digit toggles, Hide Actions): docs/dashboard-bucket-visibility.md
• Local dev, tests, demos: CONTRIBUTING.md and scripts/run_demo_scenarios.py