Without any users I find the Discourse Bitnami AWS stack to be intermittently very slow on t2.micro, where a page can take up to 15 seconds to load on multiple machines at multiple locations. What's puzzling is that I've been looking at top and it doesn't indicate anything waiting on any resources: CPU, Memory, and Disk are never overloaded.
I've even setup Elasticache for Redis which freed up ~200MB of RAM (before it was going into swap) and that hasn't helped. I've looked at the page load times and it's usually taking a long time to load the json files associated with each page, meaning it's still server-side. I always make sure each refresh I do to check is non-cached. Not sure what else to check. Any ideas? I'm new to Discourse and Ruby-on-Rails which is why I chose the Bitnami stack.
Edit: No signs of network congestion or lost packets, either. Even checked using tcpdump.
Edit2: I can't get the Bitnami stack for Discourse communicating with my Redis Elasticache. I tried removing the redis_path line in discourse.conf and replacing it with the correct redis_host line, but that didn't work. I know the Elasticache is setup right because it responds to commands from redis_cli.