4 ms·

> try rebooting the instances I haven't done anything with serverless, but surely the class of problems that would be fixed by an instance restart don't happen

• sippingjippers 6y ago
  It was intended more to invoke general ideas about management ease than being a specific remediation, however elsewhere in the thread there is an example of a diagramming tool split out across 37 individual AWS services/service instances. In a traditional design, this is conceivably something where all state and execution could easily fit in one VM, or perhaps one container with the state hoisted off to a managed service. In this case we could conceivably fix some problems with an app like that literally just by kicking the VM