In Rails 5.0 `Arel::UpdateManager.new` doesn't expect any parameters.

This commit makes the migration not pass any parameters if on Rails 5.0.

Additionally the current application setting is also updated to disable
the usage of DSA keys.

Fix missing key change in 20171124104327_migrate_kubernetes_service_to_new_clusters_architectures.rb

"Maintainer" will be freed to be used for #42751

Signed-off-by: Dmitriy Zaporozhets <dmitriy.zaporozhets@gmail.com>

Signed-off-by: Dmitriy Zaporozhets <dmitriy.zaporozhets@gmail.com>

Signed-off-by: Dmitriy Zaporozhets <dmitriy.zaporozhets@gmail.com>

Signed-off-by: Dmitriy Zaporozhets <dmitriy.zaporozhets@gmail.com>

Signed-off-by: Dmitriy Zaporozhets <dmitriy.zaporozhets@gmail.com>

Fixes that make this work:

* A change in Ruby (https://github.com/ruby/ruby/commit/ce635262f53b760284d56bb1027baebaaec175d1)
requires passing in the exact required length for OpenSSL keys and IVs.

* Ensure the secrets.yml is generated before any prepended modules are
loaded. This is done by renaming the `secret_token.rb` initializer to
`01_secret_token.rb`, which is a bit ugly but involves the least impact on
other files.

This index is added on `(target_project_id, iid)` and has a `WHERE state
= 'opened'` condition. Using this index we can drastically improve the
performance of the query used to count the total number of merge
requests in a group. Without this index the query would eventually
perform the following:

    ->  Index Scan using index_merge_requests_on_target_project_id_and_iid on merge_requests  (cost=0.43..4.89 rows=7 width=4) (actual time=0.058..0.353 rows=6 loops=228)
          Index Cond: (target_project_id = projects.id)
          Filter: ((state)::text = 'opened'::text)
          Rows Removed by Filter: 141
          Buffers: shared hit=34351 dirtied=1

Out of the ~180 milliseconds the entire query would take, around 170
milliseconds was spent in just this segment. With the index in place,
the above segment is turned into the following:

    ->  Index Only Scan using yorick_test on merge_requests  (cost=0.42..0.55 rows=7 width=4) (actual time=0.004..0.010 rows=6 loops=228)
          Index Cond: (target_project_id = projects.id)
          Heap Fetches: 419
          Buffers: shared hit=1381

The index also reduces the total query time to roughly 10 milliseconds.

Signed-off-by: Dmitriy Zaporozhets <dmitriy.zaporozhets@gmail.com>

Signed-off-by: Dmitriy Zaporozhets <dmitriy.zaporozhets@gmail.com>

We request Gitaly in a N+1 manner to build discussion diffs. Once the diffs are from different revisions, it's hard to make a single request to the service in order to build the whole response.
With this change we solve this problem and simplify a lot fetching this piece of info.