Keen · Rate Limits

Keen Rate Limits

Keen enforces per-minute rate limits per request type (events have no limit; ad-hoc queries and extractions are 200/min; updates and deletes are 10/min). The platform also applies fuzzy per-organization query concurrency limits and hard payload/response size caps. When a limit is hit all requests of that type are blocked for the next minute and return 429.

Keen Rate Limits is the machine-readable rate-limit profile for Keen on the APIs.io network, conforming to the API Commons Rate Limits specification.

It captures 13 rate-limit definitions, measuring requests_per_minute, concurrent_requests, events_scanned, events_extracted, and bytes.

The profile also includes 4 backoff/retry policies defined and response codes documented for throttled, quotaExceeded, and queryTimeout.

Tagged areas include Rate Limiting, Analytics, and Event Analytics.

13 Limits Throttle: 429 Quota: 429
Rate LimitingAnalyticsEvent Analytics

Limits

Recording events organization
requests_per_minute · minute
-1
No documented rate limit on event ingestion.
Ad-hoc queries organization
requests_per_minute · minute
200
Cached queries and datasets do not count against this limit.
Extractions organization
requests_per_minute · minute
200
Updates organization
requests_per_minute · minute
10
Deletes organization
requests_per_minute · minute
10
Delete a Collection organization
requests_per_minute · minute
100
Query concurrency organization
concurrent_requests
fuzzy per-organization concurrency, varies by workload
Returns 429 when exceeded.
Synchronous extraction scan request
events_scanned
1000000
Synchronous extractions can scan up to 1M events and return up to 100K.
Asynchronous extraction request
events_extracted
10000000
Response size request
bytes
157286400
150 MB max response per query.
Single event size request
bytes
900000
Bulk event payload size request
bytes
10000000
Query timeout request
seconds
300
Returns 504 if exceeded.

Policies

Block-and-cool
When a per-minute request-type limit is exceeded, all requests of that type are blocked for the next minute. Restrictions lift automatically once volume drops below the threshold.
Backoff
Clients should pause and retry after the one-minute window when receiving 429 responses.
Cached-first
Cached queries and cached datasets bypass the ad-hoc query rate limit and should be preferred for high-frequency dashboards.
Concurrency tuning
Concurrency limits are fuzzy and workload-dependent; clients should serialize heavy queries and batch ingestion where possible.

Sources