Development of OS update policy ideology: is it necessary and how often?

Question

JavaScript

Hello!
Steps in order:
We installed the OS from scratch (client or server - it doesn't matter now) and installed all the latest updates.
Installed and configured the necessary software.
Everything works as it should.
The next update package for the OS is coming out.
Questions:
Is it worth updating the OS constantly when the next batch of updates is released, if everything works or ...?
How do you operate?
By what criteria to decide: to be updated or not?
Do you have a plan and stick to it, or is everything as usual: the software crashed after the OS update and we begin to understand / restore performance?
In advance, Thanks everyone for the helpful replies!
PS: a question regarding ALL OS!(not only on windows, or on *nix-like ones!)

Answer 1

Depends on budget and downtime criticality. In some cases, auto-update, in some cases / their turnips, a test environment, a phased update, in some backups / snapshots / checkpoints, in some also monitoring of news about updates and what they close, what are the consequences. There are many options.
I think you should clarify what you want to update, it's ambiguous. A workstation that will crash (or crash without an update) or an elastic stack cluster where something you use will be deprecated in the new version.

Answer 2

A very difficult question, or rather a complicated answer. It is complicated not so much by difficulty as by the wealth of choice, i.e. options depending on conditions.
You can update a test machine, and ideally, more than one, then check for updates on them, then, if no jambs are found that are critical for work, “roll out” on normal machines, and ideally on groups of machines, but not all at once. For one, create a patched (updated) distribution for future deployment, or rather several, for different configurations and one - clean, without settings.
And this is the smallest thing, even for the answer can not be considered. There are several more options, but each one needs to be described based on conditions that are not reflected in the question.
And then make an image or backup for a "clean" deployment. So that you do not have to do the same operations again.

Answer 3

Is it about Windows?
Every other Tuesday, Microsoft releases security updates. (Sometimes there are emergency ones, out of turn) Most often, there is a point in putting them.
If you have a large fleet of machines and something critical depends on the "software crash", then you should test it in advance, or be able to roll back very quickly.

Answer 4

Here you are all the experts.
Tell me, how do you test updates if you have, for example, at least 200. They are assembled on, I exaggerate, but still, 5 motherboards from different manufacturers and 10 video cards? Again, they are mixed in a certain proportion. There are machines for different purposes ...
How, how do you test everything? How many people do this, even if updates are made once a month?
It was I who raised the question here about problems with hardware ... And now add problems with software that differs from PC to PC ...
How? ... I am
silent about the server. Not up to them))