Limits
Four limits apply to every caller. Two are settings; two are fixed by the system.
| Limit | Default | Changeable |
|---|---|---|
| Records in one request | 1,000 | Yes, per group |
| Size of the request body | 4,096 KB (4 MB) | No |
| Requests per minute, per group | 60 | Yes, per group |
| Writes at once, per organisation | 1 | No |
Why two are fixed
- The body size has to match the web server in front of the application. A larger figure here would be refused before BrizoConsol ever saw the request.
- One write at a time per organisation exists because a write runs inside the request
that sent it - up to a thousand records go in, in one transaction, on a worker - and two at once for
the same company would count the same month twice. A second write arriving while one is in flight is
refused with
429and codewrite_in_flight, not queued: nothing is written for it, and it is sent again once the first has finished. A read is never held behind a write.
The per-minute allowance is the group’s
The allowance is counted per group, not per token and not per credential. Minting a second token, or creating a second credential, does not buy a second sixty a minute - the group shares the one figure the limits screen holds for it. A token that cannot name its group is counted against its credential instead, and a call that names neither against the address it came from.
Both the allowance and the one-write rule are enforced on the server. A caller that keeps a month whole, one request, and waits for the answer never meets either.
Changing a group’s limits
System administration → API limits, visible to the system administrator only.
Each group on Premium is listed with its own values and whether it uses the defaults. A field left empty means “use the default”; filling one changes that limit for that group alone. The change is written to the System Audit Trail, with the value before and after.
What the caller is told
The schema endpoint states the figures a request is measured against:
GET /api/v1/schemas/trial_balance
{
"...": "...",
"record_limit": 250,
"limits": {
"records_per_request": 1000, "payload_kilobytes": 4096,
"requests_per_minute": 60, "requests_at_once": 1, "amount_tolerance": 0.01
},
"caps": {
"accounts": 300, "invoices_per_month": 1000, "bills_per_month": 1000,
"document_lines": 100, "journal_lines_per_month": 1000, "operation_rows_per_month": 1000
}
}
record_limit is the group's own when the call carries a live token, and the default when it does
not - a caller is never told one number and refused against another. limits and
caps state the installation's figures, which no group goes beyond. A request over the record
limit is refused with 422; a caller over its per-minute allowance gets 429, and so
does a write that arrives while another write for the same company is still running.
What one company may hold
The six caps on what a company may hold through the API, and what one month may carry, have their own page: Caps per company.