From 1c4e2c8e5051a079ea15ce398f39f2acaf34d0df Mon Sep 17 00:00:00 2001
From: Gavin Halliday <gavin.halliday@lexisnexis.com>
Date: Thu, 24 Oct 2024 13:09:03 +0100
Subject: [PATCH] Add summary of throughput for the set of rows

Signed-off-by: Gavin Halliday <gavin.halliday@lexisnexis.com>
---
 tools/roxie/extract-roxie-timings.py | 34 ++++++++++++++++++++++++----
 1 file changed, 30 insertions(+), 4 deletions(-)

diff --git a/tools/roxie/extract-roxie-timings.py b/tools/roxie/extract-roxie-timings.py
index b781fc815b8..b6a4e75d537 100755
--- a/tools/roxie/extract-roxie-timings.py
+++ b/tools/roxie/extract-roxie-timings.py
@@ -23,6 +23,7 @@
 import sys
 import re
 import argparse
+import datetime
 
 def calculateDerivedStats(curRow):
 
@@ -91,18 +92,19 @@ def calculateSummaryStats(curRow, numCpus, numRows):
     timeQueryResponseSeconds = float(curRow.get("elapsed", 0.0)) / 1000
     avgTimeQueryResponseSeconds = timeQueryResponseSeconds / numRows if numRows else 0
 
-    maxPerCpuTransactionsPerSecond = 1000 * numRows / timeTotalCpu if timeTotalCpu else 0
-    maxTransactionsPerSecond = numCpus * maxPerCpuTransactionsPerSecond
+    perCpuTransactionsPerSecond = 1000 * numRows / timeTotalCpu if timeTotalCpu else 0
+    maxTransactionsPerSecond = numCpus * perCpuTransactionsPerSecond
     maxWorkerThreads = numCpus / workerCpuLoad if workerCpuLoad else 0
     maxFarmers = maxTransactionsPerSecond * avgTimeQueryResponseSeconds
 
+    curRow["perCpuTransactionsPerSecond"] = perCpuTransactionsPerSecond  # recorded but not reported
     curRow["MaxTransactionsPerSecond"] = maxTransactionsPerSecond
     curRow["MaxWorkerThreads"] = maxWorkerThreads
     curRow["MaxFarmers"] = maxFarmers
 
     #Expected cpu load for 10 transactions per second per node
-    if maxPerCpuTransactionsPerSecond:
-        curRow["ExpectedCpuLoad10"] = 10 / maxPerCpuTransactionsPerSecond
+    if perCpuTransactionsPerSecond:
+        curRow["ExpectedCpuLoad10"] = 10 / perCpuTransactionsPerSecond
 
 def printRow(curRow):
 
@@ -184,6 +186,11 @@ def printRow(curRow):
                         curRow["time"] = row[i+1]
                         break
 
+                if minTimeStamp == '' or timestamp < minTimeStamp:
+                    minTimeStamp = timestamp
+                if maxTimeStamp == '' or timestamp > maxTimeStamp:
+                    maxTimeStamp = timestamp
+
                 nesting = list()
                 prefix = ''
                 suppress = 0
@@ -275,6 +282,16 @@ def printRow(curRow):
     allStats["MaxFarmers"] = 1
     allStats["ExpectedCpuLoad10"] = 1
 
+    elapsed = 0
+    try:
+        minTime = datetime.datetime.strptime(minTimeStamp, '%Y-%m-%d %H:%M:%S.%f')
+        maxTime = datetime.datetime.strptime(maxTimeStamp, '%Y-%m-%d %H:%M:%S.%f')
+        elapsed = (maxTime - minTime).seconds
+        print(f"Time range: ['{minTimeStamp}'..'{maxTimeStamp}'] = {elapsed}s")
+
+    except:
+        pass
+
     # Create a string containing all the stats that were found in the file.
     headings =  'id'
     for statName in allStats:
@@ -352,3 +369,12 @@ def sortFunc(cur):
                 printRow(centileRows[centile])
 
             print()
+            #These stats are only really revelant if it is including all the transactions from all services
+            #they may need rethinking a little.
+            if elapsed and numRows > 1:
+                perCpuTransactionsPerSecond = totalRow["perCpuTransactionsPerSecond"]
+                #elapsed is time from end of 1st transaction to end of last transaction - so subtract 1 from number of rows
+                actualTransationsPerSecond = (numRows - 1) / elapsed
+                expectedCpuLoad = actualTransationsPerSecond / perCpuTransactionsPerSecond if perCpuTransactionsPerSecond else 0
+                print(f"Transactions: Throughput={actualTransationsPerSecond}/s Time={1/actualTransationsPerSecond}s  ExpectedCpuLoad={expectedCpuLoad}")
+            print()