[
https://issues.apache.org/jira/browse/FELIX-6400?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Carsten Ziegeler updated FELIX-6400:
------------------------------------
Description:
the components health check is each time (every 5 secs) iteration over all DS
components just to check for a few. The iteration includes access to the
service registry which (in Felix framework 6.x) has a lock
There are two optimizations:
- components check should only check for the configured ones instead of
iterating over all components just to find the ones interested in
- the resultcan be cached until an event occurs (service changed event of the
SCR runtime as the runtime has a change count property to exactly avoid polling)
was:
the components readycheck is each time (every 5 secs) iteration over all DS
components just to check for a few. The iteration includes access to the
service registry which (in Felix framework 6.x) has a lock
There are two optimizations:
- components check should only check for the configured ones instead of
iterating over all components just to find the ones interested in
- the resultcan be cached until an event occurs (service changed event of the
SCR runtime as the runtime has a change count property to exactly avoid polling)
> Reduce resource consumption during component checks
> ---------------------------------------------------
>
> Key: FELIX-6400
> URL: https://issues.apache.org/jira/browse/FELIX-6400
> Project: Felix
> Issue Type: Improvement
> Components: System Ready
> Affects Versions: systemready-0.4.2
> Reporter: Carsten Ziegeler
> Priority: Major
> Fix For: systemready-0.4.3
>
>
> the components health check is each time (every 5 secs) iteration over all DS
> components just to check for a few. The iteration includes access to the
> service registry which (in Felix framework 6.x) has a lock
> There are two optimizations:
> - components check should only check for the configured ones instead of
> iterating over all components just to find the ones interested in
> - the resultcan be cached until an event occurs (service changed event of the
> SCR runtime as the runtime has a change count property to exactly avoid
> polling)
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
