[METRICS SDK] Fix hash collision in MetricAttributes

sdk/include/opentelemetry/sdk/common/attributemap_hash.h

@@ -74,27 +74,6 @@ inline size_t GetHashForAttributeMap(const OrderedAttributeMap &attribute_map)
   return seed;
 }
 
-// Calculate hash of keys and values of KeyValueIterable, filtered using callback.
-inline size_t GetHashForAttributeMap(
-    const opentelemetry::common::KeyValueIterable &attributes,
-    nostd::function_ref<bool(nostd::string_view)> is_key_present_callback)
-{
-  AttributeConverter converter;
-  size_t seed = 0UL;
-  attributes.ForEachKeyValue(
-      [&](nostd::string_view key, opentelemetry::common::AttributeValue value) noexcept {
-        if (!is_key_present_callback(key))
-        {
-          return true;
-        }
-        GetHash(seed, key);
-        auto attr_val = nostd::visit(converter, value);
-        nostd::visit(GetHashForAttributeValueVisitor(seed), attr_val);
-        return true;
-      });
-  return seed;
-}
-
 template <class T>
 inline size_t GetHash(T value)
 {

sdk/include/opentelemetry/sdk/metrics/instruments.h

@@ -65,6 +65,7 @@ struct InstrumentDescriptor
 };
 
 using MetricAttributes               = opentelemetry::sdk::metrics::FilteredOrderedAttributeMap;
+using MetricAttributesHash           = opentelemetry::sdk::metrics::FilteredOrderedAttributeMapHash;
 using AggregationTemporalitySelector = std::function<AggregationTemporality(InstrumentType)>;
 
 /*class InstrumentSelector {

sdk/include/opentelemetry/sdk/metrics/state/async_metric_storage.h

@@ -71,24 +71,22 @@ class AsyncMetricStorage : public MetricStorage, public AsyncWritableMetricStora
 
       auto aggr = DefaultAggregation::CreateAggregation(aggregation_type_, instrument_descriptor_);
       aggr->Aggregate(measurement.second);
-      auto hash = opentelemetry::sdk::common::GetHashForAttributeMap(measurement.first);
-      auto prev = cumulative_hash_map_->Get(hash);
+      auto prev = cumulative_hash_map_->Get(measurement.first);
       if (prev)
       {
         auto delta = prev->Diff(*aggr);
         // store received value in cumulative map, and the diff in delta map (to pass it to temporal
         // storage)
-        cumulative_hash_map_->Set(measurement.first, std::move(aggr), hash);
-        delta_hash_map_->Set(measurement.first, std::move(delta), hash);
+        cumulative_hash_map_->Set(measurement.first, std::move(aggr));
+        delta_hash_map_->Set(measurement.first, std::move(delta));
       }
       else
       {
         // store received value in cumulative and delta map.
         cumulative_hash_map_->Set(
             measurement.first,
-            DefaultAggregation::CloneAggregation(aggregation_type_, instrument_descriptor_, *aggr),
-            hash);
-        delta_hash_map_->Set(measurement.first, std::move(aggr), hash);
+            DefaultAggregation::CloneAggregation(aggregation_type_, instrument_descriptor_, *aggr));
+        delta_hash_map_->Set(measurement.first, std::move(aggr));
       }
     }
   }

sdk/include/opentelemetry/sdk/metrics/state/attributes_hashmap.h

@@ -15,6 +15,7 @@
 #include "opentelemetry/sdk/common/attribute_utils.h"
 #include "opentelemetry/sdk/common/attributemap_hash.h"
 #include "opentelemetry/sdk/metrics/aggregation/aggregation.h"
+#include "opentelemetry/sdk/metrics/state/filtered_ordered_attribute_map.h"
 #include "opentelemetry/sdk/metrics/view/attributes_processor.h"
 #include "opentelemetry/version.h"
 
@@ -29,9 +30,8 @@ using opentelemetry::sdk::common::OrderedAttributeMap;
 constexpr size_t kAggregationCardinalityLimit = 2000;
 const std::string kAttributesLimitOverflowKey = "otel.metrics.overflow";
 const bool kAttributesLimitOverflowValue      = true;
-const size_t kOverflowAttributesHash          = opentelemetry::sdk::common::GetHashForAttributeMap(
-    {{kAttributesLimitOverflowKey,
-               kAttributesLimitOverflowValue}});  // precalculated for optimization
+const MetricAttributes kOverflowAttributes = {
+    {kAttributesLimitOverflowKey, kAttributesLimitOverflowValue}};  // precalculated for optimization
 
 class AttributeHashGenerator
 {
@@ -42,18 +42,19 @@ class AttributeHashGenerator
   }
 };
 
-class AttributesHashMap
+template <typename CustomHash=MetricAttributesHash>
+class AttributesHashMapWithCustomHash
 {
 public:
-  AttributesHashMap(size_t attributes_limit = kAggregationCardinalityLimit)
+  AttributesHashMapWithCustomHash(size_t attributes_limit = kAggregationCardinalityLimit)
       : attributes_limit_(attributes_limit)
   {}
-  Aggregation *Get(size_t hash) const
+  Aggregation *Get(const MetricAttributes &attributes) const
   {
-    auto it = hash_map_.find(hash);
+    auto it = hash_map_.find(attributes);
     if (it != hash_map_.end())
     {
-      return it->second.second.get();
+      return it->second.get();
     }
     return nullptr;
   }
@@ -62,7 +63,7 @@ class AttributesHashMap
    * @return check if key is present in hash
    *
    */
-  bool Has(size_t hash) const { return hash_map_.find(hash) != hash_map_.end(); }
+  bool Has(const MetricAttributes& attributes) const { return hash_map_.find(attributes) != hash_map_.end(); }
 
   /**
    * @return the pointer to value for given key if present.
@@ -71,108 +72,115 @@ class AttributesHashMap
    */
   Aggregation *GetOrSetDefault(const opentelemetry::common::KeyValueIterable &attributes,
                                const AttributesProcessor *attributes_processor,
-                               std::function<std::unique_ptr<Aggregation>()> aggregation_callback,
-                               size_t hash)
+                               std::function<std::unique_ptr<Aggregation>()> aggregation_callback)
   {
-    auto it = hash_map_.find(hash);
+    MetricAttributes attr{attributes, attributes_processor};
+
+    auto it = hash_map_.find(attr);
     if (it != hash_map_.end())
     {
-      return it->second.second.get();
+      return it->second.get();
     }
 
     if (IsOverflowAttributes())
     {
       return GetOrSetOveflowAttributes(aggregation_callback);
     }
 
-    MetricAttributes attr{attributes, attributes_processor};
-
-    hash_map_[hash] = {attr, aggregation_callback()};
-    return hash_map_[hash].second.get();
+    hash_map_[attr] = aggregation_callback();
+    return hash_map_[attr].get();
   }
 
-  Aggregation *GetOrSetDefault(std::function<std::unique_ptr<Aggregation>()> aggregation_callback,
-                               size_t hash)
+  Aggregation *GetOrSetDefault(const MetricAttributes &attributes,
+                               std::function<std::unique_ptr<Aggregation>()> aggregation_callback)
   {
-    auto it = hash_map_.find(hash);
+    auto it = hash_map_.find(attributes);
     if (it != hash_map_.end())
     {
-      return it->second.second.get();
+      return it->second.get();
     }
 
     if (IsOverflowAttributes())
     {
       return GetOrSetOveflowAttributes(aggregation_callback);
     }
 
-    MetricAttributes attr{};
-    hash_map_[hash] = {attr, aggregation_callback()};
-    return hash_map_[hash].second.get();
+    hash_map_[attributes] = aggregation_callback();
+    return hash_map_[attributes].get();
   }
 
-  Aggregation *GetOrSetDefault(const MetricAttributes &attributes,
-                               std::function<std::unique_ptr<Aggregation>()> aggregation_callback,
-                               size_t hash)
+  Aggregation *GetOrSetDefault(MetricAttributes &&attributes,
+                               std::function<std::unique_ptr<Aggregation>()> aggregation_callback)
   {
-    auto it = hash_map_.find(hash);
+    auto it = hash_map_.find(attributes);
     if (it != hash_map_.end())
     {
-      return it->second.second.get();
+      return it->second.get();
     }
 
     if (IsOverflowAttributes())
     {
       return GetOrSetOveflowAttributes(aggregation_callback);
     }
 
-    MetricAttributes attr{attributes};
-
-    hash_map_[hash] = {attr, aggregation_callback()};
-    return hash_map_[hash].second.get();
+    auto result = hash_map_.emplace(std::move(attributes), aggregation_callback());
+    return result.first->second.get();
   }
-
   /**
    * Set the value for given key, overwriting the value if already present
    */
   void Set(const opentelemetry::common::KeyValueIterable &attributes,
            const AttributesProcessor *attributes_processor,
-           std::unique_ptr<Aggregation> aggr,
-           size_t hash)
+           std::unique_ptr<Aggregation> aggr)
   {
-    auto it = hash_map_.find(hash);
+    MetricAttributes attr{attributes, attributes_processor};
+
+    auto it = hash_map_.find(attr);
     if (it != hash_map_.end())
     {
-      it->second.second = std::move(aggr);
+      it->second = std::move(aggr);
     }
     else if (IsOverflowAttributes())
     {
-      hash_map_[kOverflowAttributesHash] = {
-          MetricAttributes{{kAttributesLimitOverflowKey, kAttributesLimitOverflowValue}},
-          std::move(aggr)};
+      hash_map_[kOverflowAttributes] = std::move(aggr);
     }
     else
     {
-      MetricAttributes attr{attributes, attributes_processor};
-      hash_map_[hash] = {attr, std::move(aggr)};
+      hash_map_[std::move(attr)] = std::move(aggr);
     }
   }
 
-  void Set(const MetricAttributes &attributes, std::unique_ptr<Aggregation> aggr, size_t hash)
+  void Set(const MetricAttributes &attributes, std::unique_ptr<Aggregation> aggr)
   {
-    auto it = hash_map_.find(hash);
+    auto it = hash_map_.find(attributes);
     if (it != hash_map_.end())
     {
-      it->second.second = std::move(aggr);
+      it->second = std::move(aggr);
     }
     else if (IsOverflowAttributes())
     {
-      hash_map_[kOverflowAttributesHash] = {
-          MetricAttributes{{kAttributesLimitOverflowKey, kAttributesLimitOverflowValue}},
-          std::move(aggr)};
+      hash_map_[kOverflowAttributes] = std::move(aggr);
     }
     else
     {
-      hash_map_[hash] = {attributes, std::move(aggr)};
+      hash_map_[attributes] = std::move(aggr);
+    }
+  }
+
+  void Set(MetricAttributes &&attributes, std::unique_ptr<Aggregation> aggr)
+  {
+    auto it = hash_map_.find(attributes);
+    if (it != hash_map_.end())
+    {
+      it->second = std::move(aggr);
+    }
+    else if (IsOverflowAttributes())
+    {
+      hash_map_[kOverflowAttributes] = std::move(aggr);
+    }
+    else
+    {
+      hash_map_[std::move(attributes)] = std::move(aggr);
     }
   }
 
@@ -184,7 +192,7 @@ class AttributesHashMap
   {
     for (auto &kv : hash_map_)
     {
-      if (!callback(kv.second.first, *(kv.second.second.get())))
+      if (!callback(kv.first, *(kv.second.get())))
       {
         return false;  // callback is not prepared to consume data
       }
@@ -197,8 +205,13 @@ class AttributesHashMap
    */
   size_t Size() { return hash_map_.size(); }
 
+#ifdef UNIT_TESTING
+  size_t BucketCount() { return hash_map_.bucket_count(); }
+  size_t BucketSize(size_t n) { return hash_map_.bucket_size(n); }
+#endif
+
 private:
-  std::unordered_map<size_t, std::pair<MetricAttributes, std::unique_ptr<Aggregation>>> hash_map_;
+  std::unordered_map<MetricAttributes, std::unique_ptr<Aggregation>, CustomHash> hash_map_;
   size_t attributes_limit_;
 
   Aggregation *GetOrSetOveflowAttributes(
@@ -210,19 +223,22 @@ class AttributesHashMap
 
   Aggregation *GetOrSetOveflowAttributes(std::unique_ptr<Aggregation> agg)
   {
-    auto it = hash_map_.find(kOverflowAttributesHash);
+    auto it = hash_map_.find(kOverflowAttributes);
     if (it != hash_map_.end())
     {
-      return it->second.second.get();
+      return it->second.get();
     }
 
     MetricAttributes attr{{kAttributesLimitOverflowKey, kAttributesLimitOverflowValue}};
-    hash_map_[kOverflowAttributesHash] = {attr, std::move(agg)};
-    return hash_map_[kOverflowAttributesHash].second.get();
+    hash_map_[kOverflowAttributes] = std::move(agg);
+    return hash_map_[kOverflowAttributes].get();
   }
 
   bool IsOverflowAttributes() const { return (hash_map_.size() + 1 >= attributes_limit_); }
 };
+
+using AttributesHashMap = AttributesHashMapWithCustomHash<>;
+
 }  // namespace metrics
 
 }  // namespace sdk