Optimize Banzai::Filter::RelativeLinkFilter

See merge request !4813

    A lot of git operations were being repeated, for example, to build a url
    you would ask if the path was a Tree, which would call a recursive routine
    in Gitlab::Git::Tree#where, then ask if the path was a Blob, which would
    call a recursive routine at Gitlab::Git::Blob#find, making reference to
    the same git objects several times. Now we call Rugged::Tree#path, which
    allows us to determine the type of the path in one pass.

    Some other minor improvement added, like saving commonly used references
    instead of calculating them each time.

Allow to fetch LFS from CI

## What does this MR do?

This adds support for fetching LFS object from CI jobs (mostly it's made for supporting GitLab CI).

## What is left?

- [x] Write tests covering a new authorization mechanism

cc @grzesiek @marin

See merge request !4465

Fix RangeError exceptions when referring to issues or merge requests outside of max database values

When using #XYZ in Markdown text, if XYZ exceeds the maximum value of a signed 32-bit integer, we get an exception when the Markdown render attempts to run `where(iids: XYZ)`. Introduce a method that will throw out out-of-bounds values.
    
Closes #18777

See merge request !4777

When using #XYZ in Markdown text, if XYZ exceeds the maximum value of a signed 32-bit integer, we
get an exception when the Markdown render attempts to run `where(iids: XYZ)`. Introduce a method
that will throw out out-of-bounds values.

Closes #18777

1. An exception would be raised if the filter was called with an invalid
   URI. Mainly because we weren't catching the `Addressable` exception.

2. This commit fixes it and adds a spec for the filter.

Previously we'd create a separate Metric instance for every method call
that would exceed the method call threshold. This is problematic because
it doesn't provide us with information to accurately get the _total_
execution time of a particular method. For example, if the method
"Foo#bar" was called 4 times with a runtime of ~10 milliseconds we'd end
up with 4 different Metric instances. If we were to then get the
average/95th percentile/etc of the timings this would be roughly 10
milliseconds. However, the _actual_ total time spent in this method
would be around 40 milliseconds.

To solve this problem we now create a single Metric instance per method.
This Metric instance contains the _total_ real/CPU time and the call
count for every instrumented method.