Rate limits

Bitpin limits each endpoint group separately: per user when you send a token, per IP otherwise. The client keeps a sliding window per bucket and waits before a request that would exceed it, so a well-behaved bot doesn't get 429s.

Defaults

These are the server's default limits. VIP and market-maker accounts often have higher ones. Ask Bitpin for your real limits and set them with Config(rate_limits=...) (see Configuration).

Bucket	Default	Used by
auth	3 / min	authenticate (per IP)
refresh	20 / min	token refresh
ws_info	10 / min	WebSocket token
market_public	200 / min	fetch_markets, fetch_currencies, fetch_networks
market_suspensions	2 / min	fetch_market_suspensions
tickers	300 / min	fetch_tickers, fetch_ticker
commissions	10,000 / day	fetch_trading_fees, fetch_trading_fee
candles	300 / min	fetch_ohlcv
orderbook	60 / min	fetch_order_book
matches	60 / min	fetch_trades
order_create	18,000 / hour	create_order
order_read	18,000 / hour	fetch_orders, fetch_order, open/closed orders
order_cancel	18,000 / hour	cancel_order, cancel_order_by_identifier
order_bulk_create	1,800 / hour	create_orders
order_bulk_cancel	1,800 / hour	cancel_orders, and also create_orders
fills	300 / min	fetch_my_trades
order_history	5 / sec	fetch_order_history
wallet	60 / min	fetch_balance, fetch_bank_accounts
deposit_address	60 / min	fetch_deposit_address
currency_networks	60 / min	fetch_currency_networks
deposits	60 / min	fetch_deposits
withdraws	60 / min	fetch_withdrawals, withdraw
withdraw_otp	10,000 / day	verify_withdraw_otp, resend_withdraw_otp

Bitpin counts a bulk create against the bulk-cancel budget too, and so does the client.

When the server says 429

The client raises RateLimitExceeded with retry_after (seconds) when the server gives a hint. An idempotent request (any GET) whose retry_after is 5 seconds or less is retried after waiting. Order creation and withdrawals are never retried.

from bitpin_sdk import errors

try:
    client.fetch_order_book("BTC/IRT")
except errors.RateLimitExceeded as exc:
    print("try again in", exc.retry_after, "seconds")

Several processes

The limiter counts per client object. Several processes that use the same key share one server budget, so give each a share, for example rate_limits={"order_create": "6000/hour"} for three processes.

Set enable_rate_limit=False to turn the client-side limiter off. The server still enforces its limits.