Rate Limits & Quotas#

Artery resolves quota from the authenticated request's organization plan. Each metered route is assigned a usage SKU such as api.request, markets.read, portfolio.read, or pnl.read, and that SKU can have per-minute and per-day limits.

API keys provide authentication and scopes; they do not select a separate monthly request tier. All keys in the same organization use the organization's effective plan quota.

Plan and SKU quotas#

-1 in the plan configuration means unlimited. A dash means that the plan does not set that window.

An organization can also have a subscription-level quota override. The override is merged on top of its plan per SKU and per window; this is how a custom enterprise contract can set explicit limits or mark a window unlimited.

Plans also gate features such as backtest, webhook.signing, history.maxDays, and sso. A feature gate can reject a request even when the relevant SKU still has quota remaining.

Response headers#

Metered authenticated responses carry the effective SKU quota state:

httpX-Quota-Sku:       pnl.read
X-Quota-Limit:     1000
X-Quota-Remaining: 999
X-Quota-Reset:     1780822653

X-Quota-Reset is a Unix timestamp in seconds for the reported quota window. When the effective window is unlimited, X-Quota-Limit and X-Quota-Remaining report unlimited.

429 handling#

When the effective SKU quota is exhausted:

httpHTTP/1.1 429 Too Many Requests
Retry-After:         60
X-Quota-Sku:       pnl.read
X-Quota-Limit:     30
X-Quota-Remaining: 0
X-Quota-Reset:     1780822653
 
{
  "error": {
    "code": "quota_exceeded",
    "message": "Quota exceeded for SKU 'pnl.read' (minute window: 30)",
    "requestId": "req_..."
  }
}

Branch on error.code === "quota_exceeded". Retry-After is the number of seconds to wait; X-Quota-Reset is the Unix-seconds reset timestamp.

WebSocket usage#

Delivered WebSocket messages are metered as ws.message against the organization plan. The stream gateway also limits a connection to 1,000 active subscriptions; hitting that connection cap returns:

jsonc{ "type": "error", "error": "max 1000 subs/conn" }

Upstream provider limits#

Provider limits are independent of the organization's Artery quota. These surface as code: upstream_rate_limited rather than quota_exceeded:

Artery retries upstream_rate_limited on GET calls automatically — see error handling.

See also#

• Error code reference — 429 envelope details
• Rate limits concept — how plan quotas are applied
• Authentication — API keys and scopes