cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
8519
Views
0
Helpful
4
Replies

maximum concurrent policy setting

agus.pracoyo
Level 1
Level 1

Which takes precedence "maximum concurrent connections" in listener global setting, or the one set on mail flow policy in HAT?

1 Accepted Solution

Accepted Solutions

"maximum concurrent connections" in listener are cumulative number of all connections from all sender hosts.

"maximum concurrent connections from a single IP" is maxumim number of connections from one sender host at a same time.

So if you set values:

maximum concurrent connections: 40

maximum concurrent connections from a single IP: 5

8 hosts will be able to initiate 5 concurrent connections at a same time.

 

Or if you set values:

maximum concurrent connections: 5

maximum concurrent connections from a single IP: 40

Only 1 host will be able to establish 5 connections at a same time or 5 hosts will be able to establish 1 connection at a same time.

 

Hope this helps.

View solution in original post

4 Replies 4

"maximum concurrent connections" in listener are cumulative number of all connections from all sender hosts.

"maximum concurrent connections from a single IP" is maxumim number of connections from one sender host at a same time.

So if you set values:

maximum concurrent connections: 40

maximum concurrent connections from a single IP: 5

8 hosts will be able to initiate 5 concurrent connections at a same time.

 

Or if you set values:

maximum concurrent connections: 5

maximum concurrent connections from a single IP: 40

Only 1 host will be able to establish 5 connections at a same time or 5 hosts will be able to establish 1 connection at a same time.

 

Hope this helps.

Is there a recommended value for these settings? We are using the default of 10 for 'Max Concurrent Connections From a Single IP'. And for 'Max Concurrent Connections' we use 100 (the default is 10). We are a higher ed institution and we suspect that this may be causing deferrals from some external collaborators. What indicators are there on the ESA/SMA or in the logs if a host hits the maximum? Thanks.

Hello,

 

The 'Default' values in place are the recommended; however, if you need to tweak them to fit your company needs, that is fine as well. If you're going to modify them, I would suggest doing so in smaller increments and then monitoring the System Capacity reports, as you do not want to be overloading your ESA/s.

 

In the case you are throttling hosts, you'll see something similar to this in the mail_logs :

 

Sun Sep 10 14:34:46 2017 Info: ICID 13 Receiving Failed: Connection limit exceeded

 

Thanks!

-Dennis M.

Thanks for the response. Checking our logs, I see entries like that and they do include the originating IPs. So, perhaps our customer whose connections are being rejected are not being rejected because of rate-limiting, but something else. Their email reputation isn't an issue either, so I am not clear on the cause.