Connection pooling in Apigee

[raanantharaman]

Hi,

I would like to know how connection polling is done by Apigee when it connects to target servers. Do we need to handle it ourselves explicitly using configurations etc ?

Some insight on this would be very helpful to understand the connection pooling mechanism in Apigee

Thanks

Ramnath

[#2]

Connection pooling is internally used by Apigee for backend. The concurrent connection is nothing but the use of connection pooling. At a particular time those many connections can be established.

Concurrent rate limit policy had the flexibility to change this at proxy level, but that is deprecated now.

This can be modified in message processor configuration.

[raanantharaman]

ok thanks for the info, May I know what is the configuration in message processor to handle the limit for connection pool ?

[#4]

these are some message processor properties which maintain the connections.

HTTPTransport.max.client.count -

The maximum number of connections that the

Message Processor processes at a given point

in time. It indicates the maximum number of

connections that are allowed to connect to

backend target server.

Most of the time the default value of 40000 will

suffice. If there is a specific reason to limit the

number of connections, this property can be

used for that purpose.

Known implications

Setting this value to 0 indicates unlimited

number of connections are allowed.

40000

HTTPServer.max.keepalive.clients -

The maximum number of keep alive client

connections that can be created with Message

Processor component.

This property applies to the connections made

from the router. The default value of -1

indicates that there is no limit for number of

keep alive connections coming from router.