[
https://issues.apache.org/jira/browse/SLING-11149?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17751312#comment-17751312
]
Joerg Hoh commented on SLING-11149:
-----------------------------------
[~rombert]: I wonder what you mean with improvement 1: "any cleanup operation".
In my opinion it does not make sense to log (and count) cleanup operations,
which did not perform any actual work (that means: No threadlocal was removed).
> Expose information about cleaned up leftover ThreadLocal values
> ----------------------------------------------------------------
>
> Key: SLING-11149
> URL: https://issues.apache.org/jira/browse/SLING-11149
> Project: Sling
> Issue Type: New Feature
> Components: Commons
> Reporter: Robert Munteanu
> Priority: Major
> Fix For: Commons Threads 3.3.0
>
> Time Spent: 50m
> Remaining Estimate: 0h
>
> As discussed in SLING-10831, the ThreadLocal cleanup feature is valuable as a
> safeguard for code that might lead to hard-to-diagnose memory leaks.
> At the same time, it does not let developers know when a leak has been
> prevented. We should improve this. The following improvements can be made:
> # log with a {{WARN}} level any cleanup operation. Care must be taken to not
> overflow the logs, so maybe a way of only reporting a cleanup for a certain
> key once can be used (memory-bound)
> # expose a counter that reports how many cleanup operations have been
> performed by a given thread pool
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
