Application node operating system memory (BASIC Plan) Why Is always 99%? + upstream prematurely closed connection while reading response header from upstream
Hi All. I have a basic (MENDIX CLOUD 1/1 Gigs) plan. I don't have enough clients to justify the best plan, the gap is just huge on performance, but I cannot justify it. The thing is my app is complex. And I have been trying to find ways to be on the safe side of the Application node operating system memory and every other metric. But no matter what I do, queues, cache, custom cache, removing queues, simplifications. I’m always on the 95-99% of the available system memory. And I’m even getting a crash often when showing to customers: "upstream prematurely closed connection while reading response header from upstream", nothing on the log. The app just restarts with an ugly message. I’m running out of options really. I cannot show it like that or get more clients like that. What is been considered on the APPLICATION NODE SYSTEM MEMORY? is the number of pages? microflows? nanos? How can I lower it. I only have 1 gig and I’m always at 1005-1010 MB. How can I slim that and try to make my app fit these limits? Is it tables? Is the relationships between them (I have 30-40 tables)… What is it? Also I could not find a way to debug it in eclipse (which AppId to hook to to check objects?). My app works great on dev. My problem is prod.