We are using 1 m5.xlarge production author instance. Whenever business does some bulk replication activities like publishing an entire tree with around 2000-3000 i18n keys for example, CPU spikes badly and high Garbage Collection is noticed. The author instance pretty much becomes unusable.
Has anyone seen such issues before?
Are there any OSGI configuration tunings available for bulk activities such as replication, page move, upload assets.
Are there any simple customisation of replication events or replication agents that can gracefully deal with this situation without needing to upscale the instance ?
Are there any OOTB product limitation on replication queue size ?