Configure the maximum operation rate for individual client connections and collectively for all connections associated with a client connection policy.
If the operation rate limit is exceeded, the server can either reject the operation or terminate the connection. You can define multiple rate limit values, making it possible to fine tune limits for both a long term average operation rate and short term operation bursts. For example, you can define a limit of one thousand operations per second and one million operations per day, which works out to an average of fewer than twelve operations per second, but with bursts of up to one thousand operations per second.
Specify rate limit strings as a maximum count, followed by a slash and a duration. The count portion must contain an integer and can be followed by the following multipliers:
- k (to indicate that the integer should be interpreted as thousands)
- m (to indicate that the integer should be interpreted as millions)
- g (to indicate that the integer should be interpreted as billions)
The duration portion must contain a time unit of milliseconds (ms), seconds (s), minutes (m), hours (h), days (d), or weeks (w) and can be preceded by an integer to specify a quantity for that unit.
For example, the following are valid rate limit strings:
- 1/s (no more than one operation over a one-second interval)
- 10K/5h (no more than ten thousand operations over a five-hour interval)
- 5m/2d (no more than five million operations over a two-day interval)
You can provide time units in many different formats. For example, a unit of seconds can be signified using s, sec, sect, second, and seconds.