Add performance instrumentation to key methods

lib/vmpooler/api/helpers.rb

@@ -299,17 +299,35 @@ def get_queue_metrics(pools, backend)
               total: 0
           }
 
-          queue[:requested] = get_total_across_pools_redis_scard(pools, 'vmpooler__provisioning__request', backend) + get_total_across_pools_redis_scard(pools, 'vmpooler__provisioning__processing', backend) + get_total_across_pools_redis_scard(pools, 'vmpooler__odcreate__task', backend)
-
-          queue[:pending]   = get_total_across_pools_redis_scard(pools, 'vmpooler__pending__', backend)
-          queue[:ready]     = get_total_across_pools_redis_scard(pools, 'vmpooler__ready__', backend)
-          queue[:running]   = get_total_across_pools_redis_scard(pools, 'vmpooler__running__', backend)
-          queue[:completed] = get_total_across_pools_redis_scard(pools, 'vmpooler__completed__', backend)
+          # Use a single pipeline to fetch all queue counts at once for better performance
+          results = backend.pipelined do |pipeline|
+            # Order matters - we'll use indices to extract values
+            pools.each do |pool|
+              pipeline.scard("vmpooler__provisioning__request#{pool['name']}")    # 0..n-1
+              pipeline.scard("vmpooler__provisioning__processing#{pool['name']}") # n..2n-1
+              pipeline.scard("vmpooler__odcreate__task#{pool['name']}")           # 2n..3n-1
+              pipeline.scard("vmpooler__pending__#{pool['name']}")                # 3n..4n-1
+              pipeline.scard("vmpooler__ready__#{pool['name']}")                  # 4n..5n-1
+              pipeline.scard("vmpooler__running__#{pool['name']}")                # 5n..6n-1
+              pipeline.scard("vmpooler__completed__#{pool['name']}")              # 6n..7n-1
+            end
+            pipeline.get('vmpooler__tasks__clone')           # 7n
+            pipeline.get('vmpooler__tasks__ondemandclone')   # 7n+1
+          end
 
-          queue[:cloning] = backend.get('vmpooler__tasks__clone').to_i + backend.get('vmpooler__tasks__ondemandclone').to_i
-          queue[:booting] = queue[:pending].to_i - queue[:cloning].to_i
-          queue[:booting] = 0 if queue[:booting] < 0
-          queue[:total]   = queue[:requested] + queue[:pending].to_i + queue[:ready].to_i + queue[:running].to_i + queue[:completed].to_i
+          n = pools.length
+          # Safely extract results with default to empty array if slice returns nil
+          queue[:requested] = (results[0...n] || []).sum(&:to_i) +
+                              (results[n...(2 * n)] || []).sum(&:to_i) +
+                              (results[(2 * n)...(3 * n)] || []).sum(&:to_i)
+          queue[:pending]   = (results[(3 * n)...(4 * n)] || []).sum(&:to_i)
+          queue[:ready]     = (results[(4 * n)...(5 * n)] || []).sum(&:to_i)
+          queue[:running]   = (results[(5 * n)...(6 * n)] || []).sum(&:to_i)
+          queue[:completed] = (results[(6 * n)...(7 * n)] || []).sum(&:to_i)
+          queue[:cloning]   = (results[7 * n] || 0).to_i + (results[7 * n + 1] || 0).to_i
+          queue[:booting]   = queue[:pending].to_i - queue[:cloning].to_i
+          queue[:booting]   = 0 if queue[:booting] < 0
+          queue[:total]     = queue[:requested] + queue[:pending].to_i + queue[:ready].to_i + queue[:running].to_i + queue[:completed].to_i
 
           queue
         end

lib/vmpooler/api/v3.rb

@@ -9,6 +9,20 @@ class V3 < Sinatra::Base
       api_version = '3'
       api_prefix  = "/api/v#{api_version}"
 
+      # Simple in-memory cache for status endpoint
+      # rubocop:disable Style/ClassVars
+      @@status_cache = {}
+      @@status_cache_mutex = Mutex.new
+      # rubocop:enable Style/ClassVars
+      STATUS_CACHE_TTL = 30 # seconds
+
+      # Clear cache (useful for testing)
+      def self.clear_status_cache
+        @@status_cache_mutex.synchronize do
+          @@status_cache.clear
+        end
+      end
+
       helpers do
         include Vmpooler::API::Helpers
       end
@@ -464,6 +478,32 @@ def update_clone_target(payload)
         end
       end
 
+      # Cache helper methods for status endpoint
+      def get_cached_status(cache_key)
+        @@status_cache_mutex.synchronize do
+          cached = @@status_cache[cache_key]
+          if cached && (Time.now - cached[:timestamp]) < STATUS_CACHE_TTL
+            return cached[:data]
+          end
+
+          nil
+        end
+      end
+
+      def set_cached_status(cache_key, data)
+        @@status_cache_mutex.synchronize do
+          @@status_cache[cache_key] = {
+            data: data,
+            timestamp: Time.now
+          }
+          # Cleanup old cache entries (keep only last 10 unique view combinations)
+          if @@status_cache.size > 10
+            oldest = @@status_cache.min_by { |_k, v| v[:timestamp] }
+            @@status_cache.delete(oldest[0])
+          end
+        end
+      end
+
       def sync_pool_templates
         tracer.in_span("Vmpooler::API::V3.#{__method__}") do
           pool_index = pool_index(pools)
@@ -646,6 +686,13 @@ def generate_request_id
       get "#{api_prefix}/status/?" do
         content_type :json
 
+        # Create cache key based on view parameters
+        cache_key = params[:view] ? "status_#{params[:view]}" : "status_all"
+
+        # Try to get cached response
+        cached_response = get_cached_status(cache_key)
+        return cached_response if cached_response
+
         if params[:view]
           views = params[:view].split(",")
         end
@@ -706,7 +753,12 @@ def generate_request_id
 
         result[:status][:uptime] = (Time.now - Vmpooler::API.settings.config[:uptime]).round(1) if Vmpooler::API.settings.config[:uptime]
 
-        JSON.pretty_generate(Hash[result.sort_by { |k, _v| k }])
+        response = JSON.pretty_generate(Hash[result.sort_by { |k, _v| k }])
+
+        # Cache the response
+        set_cached_status(cache_key, response)
+
+        response
       end
 
       # request statistics for specific pools by passing parameter 'pool'

lib/vmpooler/pool_manager.rb

@@ -1687,7 +1687,12 @@ def check_pool(pool,
 
           sync_pool_template(pool)
           loop do
+            start_time = Time.now
             result = _check_pool(pool, provider)
+            duration = Time.now - start_time
+
+            $metrics.gauge("vmpooler_performance.check_pool.#{pool['name']}", duration)
+            $logger.log('d', "[!] check_pool for #{pool['name']} took #{duration.round(2)}s") if duration > 5
 
             if result[:cloned_vms] > 0 || result[:checked_pending_vms] > 0 || result[:discovered_vms] > 0
               loop_delay = loop_delay_min

spec/integration/api/v3/status_spec.rb

@@ -17,6 +17,8 @@ def app()
   # https://rubydoc.info/gems/sinatra/Sinatra/Base#reset!-class_method 
   before(:each) do
     app.reset!
+    # Clear status cache to prevent test interference
+    Vmpooler::API::V3.clear_status_cache
   end
 
   describe 'status and metrics endpoints' do

spec/unit/api/helpers_spec.rb

@@ -125,8 +125,12 @@ class TestHelpers
           {'name' => 'p2'}
       ]
 
-      allow(redis).to receive(:pipelined).with(no_args).and_return [1,1]
-      allow(redis).to receive(:get).and_return(1,0)
+      # Mock returns 7*2 + 2 = 16 results (7 queue types for 2 pools + 2 global counters)
+      # For each pool: [request, processing, odcreate, pending, ready, running, completed]
+      # Plus 2 global counters: clone (1), ondemandclone (0)
+      # Results array: [1,1, 1,1, 1,1, 1,1, 1,1, 1,1, 1,1, 1, 0]
+      #                [req,  proc,  odc,   pend, rdy,  run,  comp, clone, odc]
+      allow(redis).to receive(:pipelined).with(no_args).and_return [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,0]
 
       expect(subject.get_queue_metrics(pools, redis)).to eq({requested: 6, pending: 2, cloning: 1, booting: 1, ready: 2, running: 2, completed: 2, total: 14})
     end
@@ -137,8 +141,8 @@ class TestHelpers
           {'name' => 'p2'}
       ]
 
-      allow(redis).to receive(:pipelined).with(no_args).and_return [1,1]
-      allow(redis).to receive(:get).and_return(5,0)
+      # Mock returns 7*2 + 2 = 16 results with clone=5 to cause negative booting
+      allow(redis).to receive(:pipelined).with(no_args).and_return [1,1,1,1,1,1,1,1,1,1,1,1,1,1,5,0]
 
       expect(subject.get_queue_metrics(pools, redis)).to eq({requested: 6, pending: 2, cloning: 5, booting: 0, ready: 2, running: 2, completed: 2, total: 14})
     end