add redis-timeseries

Signed-off-by: Jalal Mostafa <[email protected]>

BenchmarkTool/BenchmarkTool.csproj

@@ -2,7 +2,8 @@
 
   <PropertyGroup>
     <OutputType>Exe</OutputType>
-    <TargetFramework>net8.0</TargetFramework>
+    <TargetFramework>net7.0</TargetFramework>
+    <LangVersion>preview</LangVersion>
   </PropertyGroup>
 
   <ItemGroup>

BenchmarkTool/Config.cs

@@ -324,5 +324,10 @@ public static int GetRedisPort()
         {
             return GetInt(ConfigurationKeys.RedisPort);
         }
+
+        internal static int GetRedisConnectionsPerWorker()
+        {
+            return GetInt(ConfigurationKeys.RedisConnectionsPerWorker);
+        }
     }
 }

BenchmarkTool/ConfigurationKeys.cs

@@ -50,5 +50,6 @@ static class ConfigurationKeys
         public const string GlancesNIC = "GlancesNIC";
         public const string RedisHost = "RedisHost";
         public const string RedisPort = "RedisPort";
+        public const string RedisConnectionsPerWorker = "RedisConnectionPerWorkers";
     }
 }

BenchmarkTool/Database/RedisTimeSeriesDB.cs

@@ -17,7 +17,7 @@ namespace BenchmarkTool.Database
 {
     public class RedisTimeSeriesDB : IDatabase
     {
-        private static bool _initialized = false;
+        private static volatile bool _initialized = false;
 
         private static ConnectionMultiplexer _connection;
         private SERedis.IDatabase _redisDB;
@@ -44,7 +44,10 @@ public void Close()
             try
             {
                 if (_connection != null)
-                    _connection.Close();
+                {
+                    _connection.Dispose();
+                    _connection = null;
+                }
             }
             catch (Exception ex)
             {

Scripts/rdmabench.py

@@ -1,15 +1,24 @@
 #! /bin/python3
 import pandas as pd
 import numpy as np
+import os
+import sys
 
 def clean_results(file):
+    if not os.path.exists(file):
+        print(f'{file} does not exist!')
+        return pd.DataFrame()
     results = pd.read_csv(file)
     results['Date'] = pd.to_datetime(results['Date'], unit='ns')
     results = results.sort_values(by=['TargetDatabase', 'Date'])
     return results.set_index('Date')
 
+def set_type(file, typ):
+    results = clean_results(file)
+    results['Type'] = typ
+    return results
 
-def ingestion_rate(rmetrics='./Metrics-rdma.csv', metrics='./Metrics-sock.csv'):
+def ingestion_rate(path='.', rmetrics='Metrics-rdma.csv', metrics='Metrics-sock.csv'):
 
     def group_ingestion_rate(group):
         dates = group.index
@@ -21,14 +30,13 @@ def group_ingestion_rate(group):
         lat_sum = group['Latency'].sum() / 1e3
         return pd.Series({ 'IngestionRateAll': all_values / time, 'TotalTime': time, 'TotalPoints': all_values, 'IngestionRateMean': group['IngestionRatePoint'].mean(), 'IngestionRateBySum': all_values / lat_sum })
 
-    rdma_results = clean_results(rmetrics)
-    rdma_results['Type'] = 'RDMA'
-    sock_results = clean_results(metrics)
-    sock_results['Type'] = 'SOCK'
-
+    rdma_results = set_type(f'{path}/{rmetrics}', 'RDMA')
+    sock_results = set_type(f'{path}/{metrics}', 'SOCK')
     results = pd.concat([rdma_results, sock_results])
-    return results.groupby('Type').apply(group_ingestion_rate)
+    if len(results) == 0:
+        return pd.Series()
+    return results.groupby(['Type', 'ClientsNumber']).apply(group_ingestion_rate)
 
 if __name__ == '__main__':
-    dfs = ingestion_rate()
+    dfs = ingestion_rate(path=os.path.abspath(sys.argv[1] if len(sys.argv) > 1 else '.'))
     print(dfs.to_markdown())