Limits & versioning
Rate limits, plan allowances, and how the API changes over time.
Rate limits
Each account may make a set number of requests a minute, shared by its applications and set by its plan (listed on the Pricing page). A limited response returns 429 with a Retry-After header in seconds.
Tally is single-threaded
Plan allowances
Only a write counts: an entry TallyPrime saves. Reads — lookups, reports and dropdowns — are free, and so are failures and status checks. Past the allowance, writes return plan_limit_reached while reads keep working; nothing is charged automatically.
Times
Every timestamp is UTC, ISO 8601. The dashboard shows the same moments in your own zone.
Versioning
The version is in the path:
/api/v1/….v1is the first Balancely API.Within a version we only add — new fields on a response, new optional fields on a request, new operations. Parse defensively and ignore what you do not recognise.
A new version appears only for a change that would break an existing integration, and runs alongside the old one.
Connectors are versioned separately, so a connector release does not require reconnecting or re-entering credentials.
Deprecation
If an operation is ever retired we will announce it, keep it working for a stated period, and mark it in the operation catalogue at /api/v1/tally/operations — which is why a connector should read that list rather than hard-coding one.
Health endpoints
| Path | Purpose |
|---|---|
/healthz | Is the service process running? Answers immediately. |
/readyz | Are its dependencies healthy? Reports the database and whether the TallyPrime service is reachable. |
Neither requires authentication and neither exposes customer data. Also see Errors for how a temporary outage is reported to a running workflow.
Something missing on this page? Tell us and we will fix it.