Did anyone here end up trying the external AV route? I am still curious whether a separate scanner container on the same host would dodge the OOM issue or simply shift it.
This matches my experience precisely. The latency spikes were unmistakable once swap engaged, particularly during morning delivery bursts. I had assumed it was network-related until I checked `vmstat`.
For what it is worth, I eventually migrated to a 4GB plan on HostHatch rather than tuning further. The Altra performance at that tier has been flawless, though I recognize this sidesteps the original question rather than solving it within 2GB.