Limits

All TagoSQL limits are plan-tiered. Enterprise plans and TagoDeploy instances can have custom values; the tables below show the standard tiers.

Plan limits

Limit	Free	Starter	Scale
Stored queries	3	15	50
Versions kept per query	1	5	25
Parameters per query	3	10	50
Max joined tables	1 (no JOINs)	3	5
Result row cap	100	1,000	10,000
Fleet devices per query	10	50	100
Fleet time window	7 days	30 days	90 days
Query timeout	5s	15s	30s

Your profile's effective values are available on the profile limits endpoint (GET /profile/{id}/limits) under the sql group.

The two fleet rows govern multi-device queries (device_data_by_tag): how many devices one request may target, and how far back its mandatory time bound may reach. A tag filter matching more devices than the cap fails with a 400 telling you to narrow the tags or paginate with after_device.

The row cap

The row cap bounds how many rows a single execution can return. It never silently truncates a result. Instead:

• A query whose written LIMIT exceeds the cap is rejected when you store it: LIMIT exceeds the plan row cap (100). Lower it and paginate with OFFSET.
• A query with no LIMIT whose result turns out larger than the cap fails with Result exceeds the 100-row limit. Add LIMIT (and OFFSET to paginate). The rows that execution consumed still count against your data output, so fix the query rather than retrying it.
• A LIMIT at or below the cap behaves normally, and OFFSET is not capped.

For paging through large result sets, prefer a parameter cursor over a deep OFFSET: a deep offset makes the database scan and discard all skipped rows on every page.

Rate limits

Per-profile requests-per-minute caps:

Applies to	Free	Starter	Scale
Reading queries, versions, and the table catalog	60	100	500
Creating, editing, and deleting queries	30	75	150
Fresh executions (cache miss, test mode)	30	60	120
Cached reads	300	600	1,200

Over-limit requests return 429 with a Retry-After header. Each stored query can additionally set its own rate_limit_rpm; see Executing Queries.

Data output usage

TagoSQL executions are metered against your profile's data output allocation, the same allocation regular data reads use:

• Every execution that hits the database (a cache miss or a test-mode run) records the returned rows as output usage and is blocked when the monthly output allocation is exhausted.
• Cache hits are free: a served-from-cache result consumes no output.
• A row-cap error still counts the rows the database consumed (see The row cap).