How to organize a backup server?

Question

System administration

Good day.
I'm ashamed to ask such questions, but I don't want to reinvent the wheel.
Essence.
There is a server on which the web is spinning with a couple of dozen sites. This VPS server is located at the host. This morning a situation arose when the hoster had problems with the provider. Accordingly, all our sites were unavailable.
Website performance is extremely critical.
How can websites be made accessible?
If you organize a backup server from another host, then how should the interaction take place (how does switching take place, availability control, etc.)?

Answer 1

1 option. Raise sites from another hoster, synchronize databases through replication, specify the addresses of both servers in the DNS. Get two parallel servers, in case of failure of one - the second will continue to work ... Replication in master-master mode, you should monitor the status on both servers.
Option 2. On the second hoster, you raise sites, databases are also in a replica. In the dns, specify the address of only the first server, monitor its availability, if it fails, change the dns. DNS ttl should be kept small - the time of updating records depends on it. Well, accessibility monitoring is required.
I would choose the first option - it is easier to maintain.

Answer 2

howitmake.ru/blog/ubuntu/55.html like this, can be used for unloading or for your option.

Answer 3

People who solve these problems get from 80k per month (much from).
And so - galera, gluster, deployment system, so that the configs are the same, some balancer is smarter than "the server crashed - a third of the clients go to the forest" (the same failover IP for hetzner/linode/ovh).
And, yes, this whole canoe does not make sense if the number of nodes is less than three - there will be split-brain.