Well, yeah... on AWS, they were inherently scalable... Getting dumped was dirty pool, sure enough. I hope folks stick around for em.
@roamer_1 AWS scalability doesn't translate automatically to every application architecture. It is inherently difficult to scale out (as opposed to up) Twitter-type services because of the high cardinality of relationships between entities and the frequence of updates. It's not just like a simple web site with a single relational database back-end.
I believe that the actual last-minute improvements to Parler performance were done by disabling or doing asynchronous updates to the stats that indicate how many replies, likes, and echos each "tweet" had. For example, those last two days, Parler would always show the number of Echos as zero, even though that wasn't the case. Probably they did other things along these lines too.
Not trying to belittle you. It seems like you are an engineer with some good knowledge of the web. But super high scale services are an area that most developers never encounter. Before I landed at Microsoft, I thought I knew a lot about this topic, having authored some high scale (10K transactions per second) services in the financial space. At Microsoft, I eventually took a position managing a development team that had ownership for a significant component of the back-end of the Bing search engine. I wasn't the initial creator of this service, but I inherited it and in the process of improving it learned a great deal about this topic. We had hundreds of thousands of servers. The system was quite complex, but I would consider that an app like Twitter or Parler is more difficult in some many ways (at least has a different set of challenges) due to the cardinality and update frequency characteristics that I mentioned above.