Skip to content
ReplycoReplycoAPI

Rate limits

The Replyco API counts calls for each user and each operation. Your token decides the user, so every integration that signs in as the same user shares that user’s limits. A call over a limit is refused with 429 Too Many Requests.

How it works

Section titled “How it works”
Calls Limit for one user, per operation
Any operation 10 a second, and 60 a minute
Operations whose path ends in /paged, such as POST /openapi/thread/ticketsforreading/paged, POST /openapi/chat/paged and GET /openapi/customer/paged 10 a second, and 15 a minute
GET /openapi/thread/threadscount 10 a second, and 2 a minute

Each operation has its own count. Calling POST /openapi/chat/paged doesn’t use up what you have left on GET /openapi/customer/paged, and reading many different tickets through one operation still counts as one operation.

Signing in is the exception. POST /openapi/auth/token runs before there is a user, so it counts by IP address: 3 calls a minute from one address, including refreshes. Integrations on one server share that limit.

Each limit counts calls in a fixed window of one second or one minute. Once a window’s limit is reached, every further call in that window is refused with 429 until the window ends.

An example

Section titled “An example”

An integration that checks for new tickets every 3 seconds calls the ticket list 20 times a minute. That’s more than the 15 a minute for /paged paths, so from the 16th call each minute it gets 429. Checking every 10 seconds, with a filter that returns only the tickets you need, stays well inside the limit.

When you get a 429:

  1. Stop sending calls to that operation.
  2. Wait until the window has passed: one second for the per-second limit, one minute for the others.
  3. Try again, and then lower the rate at which you call, so that you stay under the limit.
Section titled “Related”