How to deploy from master to production with minimal effort in gitlab?

I am evaluating the possibility of production branch with Gitlab flow as given here: https://gitlab.com/gitlab-org/gitlab-ce/blob/master/doc/workflow/gitlab_flow.md#production-branch-with-gitlab-flow

Here are some points about our setup:

There are no feature branches as suggested in Gitlab Flow
All work is committed directly to master.
Each developer has his own fork which they rebase before starting off each work-item
Developers make frequent commits for each work item and make a merge request consisting many commits when they are done.
Tester pulls work from master and tests it. As a result Tester may have commits from two different merges but he may be only testing one at a time.

So lets assume I created a production branch from master a few days ago.

X1--X2--X3(master)
        |
        (prod)

After this Ann and Barry have made two merge requests which have been accepted

X1--X2--X3--A1--A1--A3-Am--B1-B2-Bm(master)
        |
        (prod)

Am and Bm represent the individual merge commits by Ann and Barry

Now tester tests and OKs Ann's changes.

Now I have to move Ann's commits to production. What would be the the neatest way? By neatest, I mean: minimum effort on part of Developer and admin. Ideally i would prefer to take Am and merge it to production, but Gitlab only gives me a cherry-pick option and not a full merge.

Like this:

X1--X2--X3--A1--A1--A3-Am--B1-B2-Bm(master)
                       |
                     (prod)

OR this

X1--X2--X3--A1--A1--A3--Am--B1-B2-Bm(master)
        |
        X3'-A1'-A1'-A3'-Am'(prod)

Now let me rewind this scenario. Tester finds a bug in Ann's code and asks for a fix. Ann does it and tester ok's the fix .

X1--X2--X3--A1--A1--A3--Am--B1--B2--Bm--A4--Am2(master)

Now I would like to take

X1--X2--X3--A1--A1--A3--Am--B1--B2--Bm--A4--Am2(master)
        |
        X3'-A1'-A1'-A3'-Am'-A4'-Am2'(prod)

Next tester OKs Barry's code so I merge Barry's changes to production

X1--X2--X3--A1--A1--A3--Am--B1--B2--Bm--A4--Am2(master)
        |
        X3'-A1'-A1'-A3'-Am'-A4'-Am2'--B1'--B2'--Bm'(prod)

Is this feasible and even practical?

What would be the best way to track pending/missing changes to production? I would like to use Gitlabs's "X commits behind master" feature to track this(eg at end of a sprint, if I'm 0 commits behind, it means I have all changes in this sprint). Or is there a better way?

Sorry if my question seems too broad but I really am confused and lost.

Appreciate any guidance that can make my life easier.

Solution

I had the same reflexion on gitlab prod workflow.
It seems difficult to implement this workflow with Gitlab CE.

I wanted to make this :

 R1-----------R2--R3
 /           /   /
A---B---C---D---E---F ← master
             \
              E---F---G ← new-feature

The fact is that Gitlab only allows the possibility to merge the whole branch master to prod or to cherrypick (bad idea).

So i decide to use a (imperfect) production workflow with tags.

R1          R2  R3   
A---B---C---D---E---F ← master
             \
              E---F---G ← new-feature

Hope it will help you.