Oh ok, thanks for the answer :)
Although i found some occasions where metrics are being exposed through JMX. I'll test them and let you know if succeed.(You seem to believe the 'no' very much thought:-P)

Regards,
Damian
On 7/13/2016 2:32 AM, Shawn McKinney wrote:
On Jul 12, 2016, at 6:22 PM, Shawn McKinney <[email protected]> wrote:

On Jul 12, 2016, at 5:51 PM, Damianos Metallidis <[email protected]> wrote:

I want to make a more coding question now and if we get an answer that would be 
great.
Let's say that i want to put a timer for how low the UserPage is about to take 
for the backend and i have an according metric for it.
Do you have any idea how to expose these metric on an JMX server from the 
project of fortress-web?
If so... :)
(At least) three ways to do this, in order of difficulty:

1. Server-side controls.  For example for slapd:
http://www.openldap.org/doc/admin24/limits.html

2. Client side, using apache ldap api settings:
https://directory.apache.org/api/gen-docs/latest/apidocs/org/apache/directory/api/ldap/model/message/SearchRequest.html#setTimeLimit-int-
This requires a simple code change to the fortress-core in order to expose a 
client-side search timeout setting via a new fortress.property.

3. Other ways by controlling the number of entries pulled back, paging, etc.  
Requires deeper changes to web app framework in its corresponding model and 
view components.

And there’s a larger question - why does the user list box always pull back 
every entry when it first loads?  That behavior is of course sub-optimal for 
prod envs (i.e. many thousands / millions users) and should be configurable.
Just realized I misread your question.  wrt measuring response times on server 
roundtrips and then exposing via jmx, the answer is ‘no’, can’t be done today.


Shawn

Reply via email to