Usage and limits

The last 24 hours, counted per application. A noisy batch job can only throttle itself; your ordering path keeps its own budget.

sandbox3 applications

Calls · 24 h24,103

Errors · 24 h974

Slowest p95957 ms

40ms

Webhook deliveries528

Endpoints

Calls by endpoint

Sorted by volume. The p95 is what your own timeouts should be set against, not the average.

API calls by endpoint over 24 hours
Endpoint	p95 ms	Calls	Errors
POST /v1/events	810	4,080	0.9%
GET /v1/merchants/{id}/offers	615	3,603	3.1%
POST /v1/orders	512	3,560	4.0%
GET /v1/orders/{id}	449	3,337	6.8%
GET /v1/merchants	784	2,589	7.0%
GET /v1/markets	357	2,423	3.8%
GET /v1/events/{id}/manifest	316	1,893	3.6%
POST /v1/catalog/offers	329	1,433	5.4%
POST /v1/orders/{id}/cancel	957	947	2.9%
GET /v1/orders	512	238	4.0%

Headroom

Limits are per application and per minute. The bar is the busiest minute of the last 24 hours.

Errors

Each of these has a fix in the body of the answer: the code, a sentence and the field that caused it.

landmark_required422Luanda, Maputo, Bissau and São Tomé need a landmark, not a street.245
idempotency_conflict409The same command_id arrived with a different body.233
merchant_closed409The store was outside its opening hours when the order arrived.190
rate_limited429Over 120 requests a minute on the catering portal key.152
scope_not_allowed_live403catalog:write is sandbox-only until the market agreement is recorded.129
amount_not_minor_units422An amount arrived with a decimal point. Send integers.25

Webhooks

Deliveries attempted per endpoint, including retries. A failing endpoint inflates this number without delivering anything.

Webhook delivery volume by endpoint
Endpoint	2xx	Attempts	Events
wh_orders_eu	100%	192	4
wh_payments	62%	240	5
wh_catalog_sync	97%	96	2