Reconsider output branch naming

[ronald-jaepel]

In GitLab by @r.jaepel on Nov 28, 2023, 11:32

From @j.schmoelder : I think we could also reconsider the naming convention of the output branches. While unlikely, a key collision is theoretically possible if two runs are started at the same time (see also #24). Instead, we could either simply use a unique ID (ideally s.t. we can use a long and a short format for referencing, just like for commit hashes) or if we go for some kind of queuing system, all runs must consecutive anyway, so we could simply give them an increasing counter.