Add storageInfo awesome memgraph function

Summary:
Added a new awesome function called `storageInfo`. This function
returns a list of key-value pairs containing the following (useful?)
information:
 * number of vertices
 * number of edges
 * average degree
 * memory usage
 * disk usage

The current implementation is in `awesome_memgraph_functions` but it will end up
as a separate clause for better access control.

Reviewers: teon.banek, mtomic, mferencevic

Reviewed By: teon.banek

Subscribers: pullbot, buda

Differential Revision: https://phabricator.memgraph.io/D1850

src/database/single_node/graph_db.hpp

@@ -16,6 +16,7 @@
 #include "storage/single_node/storage_gc.hpp"
 #include "transactions/single_node/engine.hpp"
 #include "utils/scheduler.hpp"
+#include "utils/stat.hpp"
 
 namespace database {
 
@@ -27,7 +28,7 @@ struct Stat {
 
   /// Vertex count is number of `VersionList<Vertex>` physically stored.
   std::atomic<int64_t> vertex_count{0};
-  
+
   /// Vertex count is number of `VersionList<Edge>` physically stored.
   std::atomic<int64_t> edge_count{0};
 
@@ -132,10 +133,15 @@ class GraphDb {
     if (vertex_count != 0) {
       stat_.avg_degree = 2 * static_cast<double>(edge_count) / vertex_count;
     } else {
-      stat_.avg_degree = 0; 
+      stat_.avg_degree = 0;
     }
   }
 
+  /// Returns the number of bytes used by the durability directory on disk.
+  uint64_t GetDurabilityDirDiskUsage() const {
+    return utils::GetDirDiskUsage(config_.durability_directory);
+  }
+
  protected:
   Stat stat_;
 

src/database/single_node/graph_db_accessor.cpp

@@ -12,6 +12,7 @@
 #include "storage/single_node/vertex_accessor.hpp"
 #include "utils/cast.hpp"
 #include "utils/on_scope_exit.hpp"
+#include "utils/stat.hpp"
 
 namespace database {
 
@@ -419,4 +420,20 @@ std::vector<std::string> GraphDbAccessor::IndexInfo() const {
   }
   return info;
 }
+
+std::map<std::string, std::string> GraphDbAccessor::StorageInfo() const {
+  std::map<std::string, std::string> info;
+
+  db_.RefreshStat();
+  auto &stat = db_.GetStat();
+
+  info.emplace("vertex_count", std::to_string(stat.vertex_count));
+  info.emplace("edge_count", std::to_string(stat.edge_count));
+  info.emplace("average_degree", std::to_string(stat.avg_degree));
+  info.emplace("memory_usage", std::to_string(utils::GetMemoryUsage()));
+  info.emplace("disk_usage", std::to_string(db_.GetDurabilityDirDiskUsage()));
+
+  return info;
+}
+
 }  // namespace database

src/database/single_node/graph_db_accessor.hpp

@@ -5,6 +5,7 @@
 #include <experimental/optional>
 #include <string>
 #include <vector>
+#include <map>
 
 #include <glog/logging.h>
 #include <cppitertools/filter.hpp>
@@ -604,6 +605,18 @@ class GraphDbAccessor {
   /* Returns a list of index names present in the database. */
   std::vector<std::string> IndexInfo() const;
 
+  /**
+   * Returns a map containing storage information.
+   *
+   * Inside the map, the following keys will exist:
+   *  - vertex_count
+   *  - edge_count
+   *  - average_degree
+   *  - memory_usage
+   *  - disk_usage
+   **/
+  std::map<std::string, std::string> StorageInfo() const;
+
   /**
    * Insert this vertex into corresponding label and label+property (if it
    * exists) index.

src/database/single_node_ha/graph_db.hpp

@@ -17,6 +17,7 @@
 #include "storage/single_node_ha/storage_gc.hpp"
 #include "transactions/single_node_ha/engine.hpp"
 #include "utils/scheduler.hpp"
+#include "utils/stat.hpp"
 
 namespace database {
 
@@ -140,6 +141,11 @@ class GraphDb {
     }
   }
 
+  /// Returns the number of bytes used by the durability directory on disk.
+  uint64_t GetDurabilityDirDiskUsage() const {
+    return utils::GetDirDiskUsage(config_.durability_directory);
+  }
+
  protected:
   Stat stat_;
 

src/database/single_node_ha/graph_db_accessor.cpp

@@ -12,6 +12,7 @@
 #include "storage/single_node_ha/vertex_accessor.hpp"
 #include "utils/cast.hpp"
 #include "utils/on_scope_exit.hpp"
+#include "utils/stat.hpp"
 
 namespace database {
 
@@ -418,4 +419,20 @@ std::vector<std::string> GraphDbAccessor::IndexInfo() const {
   }
   return info;
 }
+
+std::map<std::string, std::string> GraphDbAccessor::StorageInfo() const {
+  std::map<std::string, std::string> info;
+
+  db_.RefreshStat();
+  auto &stat = db_.GetStat();
+
+  info.emplace("vertex_count", std::to_string(stat.vertex_count));
+  info.emplace("edge_count", std::to_string(stat.edge_count));
+  info.emplace("average_degree", std::to_string(stat.avg_degree));
+  info.emplace("memory_usage", std::to_string(utils::GetMemoryUsage()));
+  info.emplace("disk_usage", std::to_string(db_.GetDurabilityDirDiskUsage()));
+
+  return info;
+}
+
 }  // namespace database

src/database/single_node_ha/graph_db_accessor.hpp

@@ -5,6 +5,7 @@
 #include <experimental/optional>
 #include <string>
 #include <vector>
+#include <map>
 
 #include <glog/logging.h>
 #include <cppitertools/filter.hpp>
@@ -595,6 +596,18 @@ class GraphDbAccessor {
   /* Returns a list of index names present in the database. */
   std::vector<std::string> IndexInfo() const;
 
+  /**
+   * Returns a map containing storage information.
+   *
+   * Inside the map, the following keys will exist:
+   *  - vertex_count
+   *  - edge_count
+   *  - average_degree
+   *  - memory_usage
+   *  - disk_usage
+   **/
+  std::map<std::string, std::string> StorageInfo() const;
+
   /**
    * Insert this vertex into corresponding label and label+property (if it
    * exists) index.

src/query/interpret/awesome_memgraph_functions.cpp

@@ -703,6 +703,17 @@ TypedValue WorkerId(TypedValue *args, int64_t nargs, const EvaluationContext &,
 }
 #endif
 
+#if defined(MG_SINGLE_NODE) || defined(MG_SINGLE_NODE_HA)
+TypedValue StorageInfo(TypedValue *, int64_t nargs, const EvaluationContext &,
+                       database::GraphDbAccessor *dba) {
+  if (nargs != 0)
+    throw QueryRuntimeException("'storageInfo' requires no arguments.");
+
+  auto info = dba->StorageInfo();
+  return std::map<std::string, TypedValue>(info.begin(), info.end());
+}
+#endif
+
 TypedValue Id(TypedValue *args, int64_t nargs, const EvaluationContext &,
               database::GraphDbAccessor *dba) {
   if (nargs != 1) {
@@ -985,6 +996,9 @@ NameToFunction(const std::string &function_name) {
 #ifdef MG_DISTRIBUTED
   if (function_name == "WORKERID") return WorkerId;
 #endif
+#if defined(MG_SINGLE_NODE) || defined(MG_SINGLE_NODE_HA)
+  if (function_name == "STORAGEINFO") return StorageInfo;
+#endif
 
   return nullptr;
 }

src/telemetry/collectors.cpp

@@ -8,6 +8,7 @@
 #include <unistd.h>
 
 #include "utils/file.hpp"
+#include "utils/stat.hpp"
 #include "utils/string.hpp"
 
 namespace telemetry {
@@ -74,17 +75,7 @@ const nlohmann::json GetResourceUsage() {
   auto cpu_total = GetCpuUsage(pid);
   cpu["usage"] = cpu_total.second;
 
-  // Parse memory usage.
-  uint64_t memory = 0;
-  auto statm_data = utils::ReadLines(fmt::format("/proc/{}/statm", pid));
-  if (statm_data.size() >= 1) {
-    auto split = utils::Split(statm_data[0]);
-    if (split.size() >= 2) {
-      memory = std::stoull(split[1]) * sysconf(_SC_PAGESIZE);
-    }
-  }
-
-  return {{"cpu", cpu}, {"memory", memory}};
+  return {{"cpu", cpu}, {"memory", utils::GetMemoryUsage()}};
 }
 
 }  // namespace telemetry

src/utils/stat.hpp

@@ -0,0 +1,45 @@
+#pragma once
+
+#include <experimental/filesystem>
+
+#include <unistd.h>
+
+#include "utils/file.hpp"
+#include "utils/string.hpp"
+
+namespace utils {
+
+/// Returns the number of bytes a directory is using on disk. If the given path
+/// isn't a directory, zero will be returned.
+inline uint64_t GetDirDiskUsage(
+    const std::experimental::filesystem::path &path) {
+  if (!std::experimental::filesystem::is_directory(path)) return 0;
+
+  uint64_t size = 0;
+  for (auto &p : std::experimental::filesystem::directory_iterator(path)) {
+    if (std::experimental::filesystem::is_directory(p)) {
+      size += GetDirDiskUsage(p);
+    } else if (std::experimental::filesystem::is_regular_file(p)) {
+      size += std::experimental::filesystem::file_size(p);
+    }
+  }
+
+  return size;
+}
+
+/// Returns the number of bytes the process is using in the memory.
+inline uint64_t GetMemoryUsage() {
+  // Get PID of entire process.
+  pid_t pid = getpid();
+  uint64_t memory = 0;
+  auto statm_data = utils::ReadLines(fmt::format("/proc/{}/statm", pid));
+  if (statm_data.size() >= 1) {
+    auto split = utils::Split(statm_data[0]);
+    if (split.size() >= 2) {
+      memory = std::stoull(split[1]) * sysconf(_SC_PAGESIZE);
+    }
+  }
+  return memory;
+}
+
+}  // namespace utils