1.We’re internet hosting a number of WordPress web sites on Aurora database. We develop the websites in a subdomain and when go dwell we simply rename the subdomain to the precise area and make the required adjustments within the database. However, each time we launch the websites, we get a variety of database connections (we run the present full processlist command) and our Aurora database server goes down. After we change the area identify to another area, it really works effective. After we disable the Elementor (WordPress rag and drop builder) builder’s header template, it really works effective. When it is nonetheless a subdomain it doesn’t trigger any points. Question monitor don’t present any suspicious queries and “present full processlist” present 1000+ queries in both “sleep” or “cleansing up” or “cleaned up” state however doesn’t kill them.
2.We now have an ec2 picture (It even have data so as to add the EFS quantity in a predefined path when it boots) that we use to create additonal ec2 cases and add these cases to load balancer group manually. Can we use that picture (12 GB) with AWS autoscaling and improve/lower ec2 cases relying on server load ? How lengthy it would take for a brand new occasion as well up if server load will increase ? Can we activate/off the autoscaling with out affecting the bottom set of ec2 cases ?
3.We now have 2 ec2 cases working 50 WP Single websites and multisites with horizontal scaling. Our structure entails load balancer, redis cache, Aurora DB (2 cpu to 4 cpu scaling) and EFS (uploads). What ec2 occasion sort you intend and why ? Ought to we use excessive ec2 varieties like M sequence or ought to we use t2/t3 occasion varieties however improve the ec2 cases from 2 to 4 (horizontal scaling) ?
