Restrict CPU usage checker to monitor sampleAndAggregate cpu usge only

As sampleAndAggregate happens every sampling_interval, our cpu usage checker
estimate the cpu usage by taking the cpu time spent in sampleAndAggregate
divided by sampling_interval.

This change will not be merged until we also have a checker monitoring
the cpu usage for submitProfile and refreshConfig.

Author: Colman Yau

codeguru_profiler_agent/profiler_disabler.py

@@ -39,12 +39,12 @@ def __init__(self, timer):
         self.timer = timer
 
     def is_cpu_usage_limit_reached(self, profile=None):
-        profiler_metric = self.timer.metrics.get("runProfiler")
-        if not profiler_metric or profiler_metric.counter < MINIMUM_MEASURES_IN_DURATION_METRICS:
+        sample_and_aggregate_metric = self.timer.metrics.get("sampleAndAggregate")
+        if not sample_and_aggregate_metric or sample_and_aggregate_metric.counter < MINIMUM_MEASURES_IN_DURATION_METRICS:
             return False
 
         sampling_interval_seconds = self._get_average_sampling_interval_seconds(profile)
-        used_time_percentage = 100 * profiler_metric.average() / sampling_interval_seconds
+        used_time_percentage = 100 * sample_and_aggregate_metric.average() / sampling_interval_seconds
 
         if used_time_percentage >= AgentConfiguration.get().cpu_limit_percentage:
             logger.debug(self.timer.metrics)

codeguru_profiler_agent/profiler_runner.py

@@ -89,10 +89,14 @@ def _run_profiler(self):
             if self.collector.flush():
                 self.is_profiling_in_progress = False
                 return True
-            sample = self.sampler.sample()
-            self.collector.add(sample)
+            self._sample_and_aggregate()
         return True
 
+    @with_timer("sampleAndAggregate")
+    def _sample_and_aggregate(self):
+        sample = self.sampler.sample()
+        self.collector.add(sample)
+
     def is_running(self):
         return self.scheduler.is_running()
 

test/acceptance/test_cpu_limit.py

@@ -31,6 +31,6 @@ def test_profiler_terminates(self):
             # With sampling_interval to be 0.01 seconds, having runProfiler as 0.5 seconds should breach
             # the cpu limit. We need to sample 20 times before we check the CPU limit
             for i in range(20):
-                self.timer.record('runProfiler', 0.5)
+                self.timer.record('sampleAndAggregate', 0.5)
 
             assert wait_for(lambda: not self.profiler.is_running(), timeout_seconds=5)

test/unit/test_profiler_disabler.py

@@ -150,7 +150,7 @@ def before(self):
         self.timer = Timer()
         self.profile = Mock(spec=Profile)
         for i in range(20):
-            self.timer.record('runProfiler', 0.5)
+            self.timer.record('sampleAndAggregate', 0.5)
         set_agent_config(sampling_interval_seconds=1, cpu_limit_percentage=10)
         self.process_duration_check = CpuUsageCheck(self.timer)
 
@@ -222,7 +222,7 @@ def before(self):
     def test_it_returns_false(self):
         self.timer.reset()
         for i in range(4):
-            self.timer.record('runProfiler', 0.5)
+            self.timer.record('sampleAndAggregate', 0.5)
         assert not self.process_duration_check.is_cpu_usage_limit_reached()