Improve CI

[solnic]

• Exlude rubocop from CI's bundle - this was needed because we want to test things when rack is not installed too and unfortunately rubocop-rails depends on rack, so rack was always installed
• Measure coverage on latest Ruby 3.3 - we're gonna have specs that run only under 3.3 once Vernier branch is merged and w/o this tweak coverage would go down
• Add rack 3.1 and redis 5.3 to the test matrix

#skip-changelog

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 98.67%. Comparing base (9d23b6f) to head (7753850).
Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2388      +/-   ##
==========================================
- Coverage   98.69%   98.67%   -0.02%     
==========================================
  Files         210      210              
  Lines       13910    13910              
==========================================
- Hits        13728    13726       -2     
- Misses        182      184       +2     

Components	Coverage Δ
sentry-ruby	99.05% <100.00%> (-0.02%)	⬇️
sentry-rails	97.34% <ø> (ø)
sentry-sidekiq	97.01% <ø> (ø)
sentry-resque	96.79% <ø> (ø)
sentry-delayed_job	98.92% <ø> (ø)
sentry-opentelemetry	100.00% <ø> (ø)

Files	Coverage Δ
...y-ruby/spec/sentry/rack/capture_exceptions_spec.rb	100.00% <100.00%> (ø)

... and 3 files with indirect coverage changes

[solnic]

This is not beautiful but it was needed as this constant may not be defined if rack is not available

[solnic]

The diff in cov against master is most likely caused by the fact that this PR runs coverage under 3.3 and master 3.2, so we can just ignore it