Details
-
Task
-
Resolution: Unresolved
-
Neutral
-
None
-
5.9.4
-
None
-
None
-
-
Empty show more show less
-
Empty show more show less
-
Yes
Description
CacheMonitor mbean exposes various metrics that can be scraped by the monitoring systems such as Prometheus or DataDog. Scraping a couple of those metrics though can have some serious implications on the system's memory allocation footprint and/or performance in general:
- info.magnolia.module.cache.mbean.CacheMonitor#getCachedKeysCount
- info.magnolia.module.cache.mbean.CacheMonitor#getCachedUUIDsCount
Both of the above perform some aggregate calculations over the entries stored in cache and may accidentally cause a multitude of those to be loaded in memory just for the iterative counting purposes.
Scaping those metrics should be done carefully and our documentation needs to address that. (or we can eliminate them completely, or somehow improve their performance which I doubt)
Checklists
Acceptance criteria