What is the DevOps Best Practices for Hybrid environments?

Question

I want to learn best practices, and I hope you could help me with your expertise.

In a large organization that uses cloud services and taking advantage of Infrastructure as Code (IaC). My two questions are as follows:

1) What is the best way to patch a cloud IaaS or hybrid environment? I s there a way to do so without downtime at all? Please explain.

2) With the powerfulness of IaC and that we can provision environments in seconds. Is there even a point to patch? Why don’t we just keep provisioning whenever we want and decommission the older ones?

Any inputs will be greatly appreciated!

Answer 1

There may be a significant speed difference between provisioning an environment and bringing up that environment. Especially when it comes to what you mentioned: IaaS and hybrid environments - those are VMs and bare metal servers - it may take many seconds/minutes for some of them to come up.

So there may still be value in patching those environments, provided, of course, that the respective patches can be done without downtime. That depends on the patches - upgrading the kernel on a linux server, for example, would require a reboot to take effect. Upgrading Python won't.

The intended usage for particular IaaS/hybrid environments is also an important aspect in this discussion. The downtime could be completely insignificant for a distributed app designed to tolerate restarts of individual servers in the pool it's running on but probably a big deal for a centralized server app serving thousands of users.