Use sharding, run across 4 machines

[hannes-mk]

No description provided.

[hannes-mk]

@HPrinz this PR is what I tried wrt. playwright sharding. Unfortunately I was not successful, so I pushed this to a later point in time.

Here's what I did

• Setting up a "matrix" for 4 "shards"
• Using the numbers in the --shard parameter of playwright (as in the documentation)

However, it did not reach faster e2e speeds: sharded and unsharded were ~3:30min. I suspect that the "splitting" of our tests did not work as expected.

Also, the PR was also incomplete wrt.

• Uploading the individual shard test reports
• consolidating the log before pushing to ris-reports

[HPrinz]

Thank you! I was hoping this would be much easier :/