From 0f5b25cefcdaab4547b01a6f150ff7a4faf52c5c Mon Sep 17 00:00:00 2001
From: ztao <t@taozj.org>
Date: Fri, 9 Sep 2022 08:32:08 -0400
Subject: [PATCH] Sample the summary observed values

---
 core/include/prometheus/summary.h | 14 +++++++++++---
 core/src/summary.cc               | 23 ++++++++++++++++++-----
 2 files changed, 29 insertions(+), 8 deletions(-)

diff --git a/core/include/prometheus/summary.h b/core/include/prometheus/summary.h
index 7035459b..aa705beb 100644
--- a/core/include/prometheus/summary.h
+++ b/core/include/prometheus/summary.h
@@ -71,14 +71,21 @@ class PROMETHEUS_CPP_CORE_EXPORT Summary {
   /// and how smooth the time window is moved. With only one age bucket it
   /// effectively results in a complete reset of the summary each time max_age
   /// has passed. The default value is 5.
+  ///
+  /// \param sample_ratio Set the sample ratio for the observed values which are
+  /// used to calculate quantiles. When used in high throughput situations, if
+  /// not sampled the SDK will cause serious performance problem. The value
+  /// valid in [0, 1], and sample_ratio = 0 or 1 will not cause sample. The
+  /// default value is 0.
   explicit Summary(const Quantiles& quantiles,
                    std::chrono::milliseconds max_age = std::chrono::seconds{60},
-                   int age_buckets = 5);
+                   int age_buckets = 5, double sample_ratio = 0);
 
-  /// \copydoc Summary::Summary(const Quantiles&,std::chrono::milliseconds,int)
+  /// \copydoc Summary::Summary(const
+  /// Quantiles&,std::chrono::milliseconds,int,double)
   explicit Summary(Quantiles&& quantiles,
                    std::chrono::milliseconds max_age = std::chrono::seconds{60},
-                   int age_buckets = 5);
+                   int age_buckets = 5, double sample_ratio = 0);
 
   /// \brief Observe the given amount.
   void Observe(double value);
@@ -94,6 +101,7 @@ class PROMETHEUS_CPP_CORE_EXPORT Summary {
   std::uint64_t count_{};
   double sum_{};
   detail::TimeWindowQuantiles quantile_values_;
+  int sample_ratio_threshold_{};
 };
 
 /// \brief Return a builder to configure and register a Summary metric.
diff --git a/core/src/summary.cc b/core/src/summary.cc
index 29902140..bb05e805 100644
--- a/core/src/summary.cc
+++ b/core/src/summary.cc
@@ -5,21 +5,34 @@
 namespace prometheus {
 
 Summary::Summary(const Quantiles& quantiles,
-                 const std::chrono::milliseconds max_age, const int age_buckets)
+                 const std::chrono::milliseconds max_age, const int age_buckets,
+                 const double sample_ratio)
     : quantiles_{quantiles},
-      quantile_values_{quantiles_, max_age, age_buckets} {}
+      quantile_values_{quantiles_, max_age, age_buckets} {
+  if (sample_ratio > 0 && sample_ratio < 1) {
+    sample_ratio_threshold_ = sample_ratio * RAND_MAX;
+  }
+}
 
 Summary::Summary(Quantiles&& quantiles, const std::chrono::milliseconds max_age,
-                 const int age_buckets)
+                 const int age_buckets, const double sample_ratio)
     : quantiles_{std::move(quantiles)},
-      quantile_values_{quantiles_, max_age, age_buckets} {}
+      quantile_values_{quantiles_, max_age, age_buckets} {
+  if (sample_ratio > 0 && sample_ratio < 1) {
+    sample_ratio_threshold_ = sample_ratio * RAND_MAX;
+  }
+}
 
 void Summary::Observe(const double value) {
+  bool should_observe =
+      (sample_ratio_threshold_ == 0 || sample_ratio_threshold_ < std::rand());
   std::lock_guard<std::mutex> lock(mutex_);
 
   count_ += 1;
   sum_ += value;
-  quantile_values_.insert(value);
+  if (should_observe) {
+    quantile_values_.insert(value);
+  }
 }
 
 ClientMetric Summary::Collect() const {