"Parallelize" the dashboard build

[CasperWA]

It takes more than half an hour to build the dashboard currently. And this is without some expected providers to come online.
One way to minimize the runtime for the build would be to "parallelize" it, if possible.

[ml-evs]

I added some simple multiprocessing-based parallelism (over providers) in #82, but unfortunately github actions only gives you 2 cores to play with so there is limited benefit (wish I had checked this first...).

On my machine it brings the dashboard build down to ~10 minutes. Another approach would be to split the dashboard build into multiple CI jobs that read from a static providers file (would need to submodule the providers repo), but this would be much more work.

[ml-evs]

Since the network should be the bottleneck, I am now testing whether overloading the 2 cores with 8 processes still leads to a reasonable speed-up.

Another alternative to the above is adding parallelism/async to the validator itself.