Each blockchain protocol and consensus mechanism has its own crash or fault tolerance. Generally, the more decentralized nodes that there are in a network, the less prone to disaster the network is.
Should all computing resources in one location that the network is deployed in fail, as long as nodes of the network are running in other locations, and a sufficient number of those nodes have the right role (e.g. validator node), the overall network will not fail.
In the unlikely event of all cloud nodes in all regions failing due to a global outage, Chainstack will work with each cloud provider to restore the services with immediate urgency.