Remove fallback to project.deployment_platform

This improves query performance of
MergeRequestsController#ci_environments_status a lot.

However this means old deployments that deployed to kubernetes clusters
with prometheus installations will no longer show performance metrics as
we cannot backfill cluster_id from deployment_platform with certainty
(clusters may be edited/added/deleted, which changes the results of
deployment_platform).

app/models/deployment.rb

@@ -206,13 +206,8 @@ class Deployment < ApplicationRecord
     end
   end
 
-  # TODO remove fallback case to deployment_platform_cluster.
-  # Otherwise we will continue to pay the performance penalty described in
-  # https://gitlab.com/gitlab-org/gitlab-ce/issues/63475
   def cluster_prometheus
-    cluster_with_fallback = cluster || deployment_platform_cluster
-
-    cluster_with_fallback.application_prometheus if cluster_with_fallback&.application_prometheus_available?
+    cluster.application_prometheus if cluster&.application_prometheus_available?
   end
 
   def ref_path

changelogs/unreleased/63475-fix-n-1.yml

@@ -0,0 +1,5 @@
+---
+title: Improve performance of MergeRequestsController#ci_environment_status endpoint
+merge_request: 30224
+author:
+type: performance

spec/models/deployment_spec.rb

@@ -341,18 +341,6 @@ describe Deployment do
 
         it { is_expected.to be_truthy }
       end
-
-      context 'fallback deployment platform' do
-        let(:cluster) { create(:cluster, :provided_by_user, environment_scope: '*', projects: [deployment.project]) }
-        let!(:prometheus) { create(:clusters_applications_prometheus, :installed, cluster: cluster) }
-
-        before do
-          expect(deployment.project).to receive(:deployment_platform).and_return(cluster.platform)
-          expect(cluster.application_prometheus).to receive(:can_query?).and_return(true)
-        end
-
-        it { is_expected.to be_truthy }
-      end
     end
   end