[OpenSIPS-Users] New usage for the ratelimit rl_check function

[xvelasco]

Hi Razvan,
I have implemented the ratelimit module using the examples you gave to me.
So, now the cps limit per gateway is being created/updated at runtime and
I'm using TAILDROP as the default algorithm. Besides the timer_interval is
set to 1.

However, I have a new issue now. For example, I have a gateway "gw_0" that
has a ratelimit set to 100. I'm generating 130 calls per second, that need
to be processed using gw_0, the call duration for each call is 2 seconds.
Once that limit is reached, the opensips should say: /Pipe gw_0 counter:101
load:0 limit:100 should be blocked/. This is the expected behavior. 

But the thing is that every second the ratelimit per gateway is reseted,
/DBG:ratelimit:rl_timer: Pipe "gw_0" load updated to 0/, and the counter for
an specific gateway starts again at 0. This produces, for example, that when
opensips is processing let's say the call 50 and it resets the ratelimit for
gw_0, the counter for that gateway is now 0 and the remaining 80 calls will
be successfully processed, due the gateway limit is 100. This behavior is
not what I'm expecting, since at the same second the gateway will be
processing more calls than the CPS we set. 

Am I missunderstanding the ratelimit module behavior?

Thanks for your help!



--
View this message in context: http://opensips-open-sip-server.1449251.n2.nabble.com/New-usage-for-the-ratelimit-rl-check-function-tp7586630p7587028.html
Sent from the OpenSIPS - Users mailing list archive at Nabble.com.