It's VM and it can be down for no reason, you need redundancy and auto scaling policy to ensure a new instance will be started when you have less than configured running, and start more instance when load is beyond threshold.
【在 g*****g 的大作中提到】 : It's VM and it can be down for no reason, you need redundancy and auto : scaling policy to ensure a new instance will be started when you have less : than configured running, and start more instance when load is beyond : threshold. : : outage
b*s
21 楼
aws也是可以选加cpu的,不过贵
【在 d*******r 的大作中提到】 : 我在国内的技术论坛上看,也是这个观点,貌似 Digital Ocean 的CPU不行,不过它家 : 比 Linode 还便宜。 : 不过想国内也能访问的话,还是得上 Linode, 有 tokyo 节点,不过据说最近国内用的 : 人有点多了,tokyo 节点也慢了下来。不知道还有哪家的服务性价比比 Linode 和 : Digital Ocean 更好。
I don't think changing from small to medium makes sense, since it will involves downtime, but certainly auto scaling can scale up/down instances by load or by time. You can specify min/max/desired too. For DB, you can use simpleDB, RDS etc. We also use in instance Cassandra cluster. Webpages can be replicated on all instances.