Merge branch '65278-fix-puma-master-counter-wipe' into 'master'

Fix active metric files being wiped after the app starts

See merge request gitlab-org/gitlab-ce!31668

Author: stanhu

changelogs/unreleased/65278-fix-puma-master-counter-wipe.yml

@@ -0,0 +1,5 @@
+---
+title: Fix active metric files being wiped after the app starts
+merge_request: 31668
+author:
+type: fixed

config.ru

@@ -17,7 +17,25 @@ end
 
 require ::File.expand_path('../config/environment', __FILE__)
 
+# The following is necessary to ensure stale Prometheus metrics don't accumulate over time.
+# It needs to be done as early as here to ensure metrics files aren't deleted.
+# After we hit our app in `warmup`, first metrics and corresponding files already being created,
+# for example in `lib/gitlab/metrics/requests_rack_middleware.rb`.
+def cleanup_prometheus_multiproc_dir
+  if dir = ::Prometheus::Client.configuration.multiprocess_files_dir
+    old_metrics = Dir[File.join(dir, '*.db')]
+
+    FileUtils.rm_rf(old_metrics)
+  end
+end
+
+def master_process?
+  Prometheus::PidProvider.worker_id.in? %w(unicorn_master puma_master)
+end
+
 warmup do |app|
+  cleanup_prometheus_multiproc_dir if master_process?
+
   client = Rack::MockRequest.new(app)
   client.get('/')
 end

config/initializers/7_prometheus_metrics.rb

@@ -51,22 +51,3 @@ def prometheus_default_multiproc_dir
     end
   end
 end
-
-def cleanup_prometheus_multiproc_dir
-  # The following is necessary to ensure stale Prometheus metrics don't
-  # accumulate over time. It needs to be done in this hook as opposed to
-  # inside an init script to ensure metrics files aren't deleted after new
-  # unicorn workers start after a SIGUSR2 is received.
-  if dir = ::Prometheus::Client.configuration.multiprocess_files_dir
-    old_metrics = Dir[File.join(dir, '*.db')]
-    FileUtils.rm_rf(old_metrics)
-  end
-end
-
-Gitlab::Cluster::LifecycleEvents.on_master_start do
-  cleanup_prometheus_multiproc_dir
-end
-
-Gitlab::Cluster::LifecycleEvents.on_master_restart do
-  cleanup_prometheus_multiproc_dir
-end

lib/gitlab/cluster/puma_worker_killer_observer.rb

@@ -15,9 +15,7 @@ def callback
       private
 
       def log_termination(worker)
-        labels = { worker: "worker_#{worker.index}" }
-
-        @counter.increment(labels)
+        @counter.increment
       end
     end
   end

spec/lib/gitlab/cluster/puma_worker_killer_observer_spec.rb

@@ -17,9 +17,7 @@
     it 'increments timeout counter' do
       worker = double(index: 0)
 
-      expect(counter)
-        .to receive(:increment)
-        .with({ worker: 'worker_0' })
+      expect(counter).to receive(:increment)
 
       subject.callback.call(worker)
     end