New post
Question
· May 25, 2022

Cache MONLBL Response Time question

#Caché

Hello to all my fellow Cache Experts.   Quick question about the MONLBL output that identifies performance issues.  There is a metric value output by MONLBL called "Obtain [NEW] connection to Cache".   Supposedly that metric captures the time taken between the request reaching the CSP Gateway and a connection to Cache being reserved for the purpose of servicing the request.   All of that being said, if that metric value is high for many requests, such as 19.x seconds, then can anyone tell me what that metric really means and how to improve (lower) the new connection to Cache time.   The company I work for is trying to analyze a few areas of the application that seems to be slower than expected and this is one of the metrics from MONLBL that stood out.   So if anyone has any ideas or suggestions about possible configuration changes or tweaks at the CSP Gateway, OS, or Cache Instance servicing the request, so a new web request inbound to Cache Namespace is near instantaneous versus taking 19.x seconds, that would be great.

Much thanks in advance for any ideas or suggestions.   GO TEAM!!!!

Product version: Caché 2017.1
Discussion (3)1
Log in or sign up to continue
Rich Taylor · Jun 8, 2022

One thought that occurs to me is to check the 'Maximum Server Connections' settings in the CSP  Gateway under Server Access.  This is the maximum number of connections allows and when reached further requests are queued until a connection is available or the client times out.  This is often used with REST services to avoid clients receiving a licensing error which is not desirable or to avoid any CSP calls from consuming the entire system.   If this is the case you should examine your license usage and CSP Sessions to see what might be overloading the server. 

Of course if this is a very active system then you may need to increase the maximum connections allowed and possibly the license as well.

Hope that helps.

1 0
Rich Taylor · Jun 15, 2022

There is no general rule-of-thumb.  It really depends on the nature of your application and usage patterns.  For example if this is primarily a REST based service I would set the maximum to your license capacity minus a reserve for administrative functions, background tasks, and reporting.

In a mixed environment you need to have a more balanced setup of REST (web) calls versus direct connections.

As I indicated you really need to have an idea of the use patterns of your application.

1 0