The things that make the metric system superior to prior systems are:

• Using a uniform set of prefixes to designate multiples and divisions of the base units.

Having one unit of say, volume (the liter), and then using prefixes when we need smaller or larger units is way better than having cups, pints, quarts, gallons, pecks, and many more.

• Having those prefixes mean powers of 10. That fits in well with our use of decimal arithmetic.

It is the first one that is most important.

For temperature there's nothing actually 'Metric-y' about Celsius (or Kelvin), because in most cases people don't use multiples or divisions of the base unit. This includes in science and engineering. An astronomer would say (and write in their paper) that a star has a temperature of 7000 K, not 7 kiloK. They would say a neutron star has a core temperature of 100 trillion K, not 100 TK or 100 teraK.

At the low end there is more use of prefixes. The scientists that work near absolute 0 do often use millikelvin and microkelvin. They also often don't. Both 10^-2 Kelvin and 10 mK would usually be acceptable.

A metric system with the same meter, liter, and gram as the current one but that had picked F and R instead of C and K would work fine and be just as 'Metric-y' as the current metric system.