Detect failures in benchmarks

[Akirathan]

In https://github.com/enso-org/enso/actions/runs/10533928803/job/29190752771, all the benchmarks fail and yet the whole job succeeded. We should ensure that the whole job exits with non-zero exit code if one of the benchmarks failed, but still uploads the results for the successful ones.

### Tasks
- [ ] Test `sbt std-benchmarks/run` with dry-run env var