Adjusting regular expressions for features extraction

[tomchop]

Some changes to the regexes in features.yaml, mostly relaxing them:

• Removed the header check for sshd/pid, given that there were two versions: \[sshd, pid: \d+\] and \[sshd\] \[\d+\]
• Removed the .* at the start of some regexes which must've made them much slower.

[jkppr]

Can you please decouple the changes to the features.yaml from the changes to the yeti analyzer?
Have this PR just focused on the features changes and add the changes for the analyzer in your #2930 PR

[tomchop]

I sent it too soon because I'm working out of the same branch, but it should get decoupled as soon as the other one is merged. :)