PTF - Performance Instance Resources Optimization - MCPT

Overview

  • The primary objective of testing was to evaluate the performance of the Baseline MCPT Environment configuration while attempting to optimize costs by adjusting instance types and reducing the number of instances. The tests were designed to compare the performance outcomes across different configurations, including variations in instance types and counts within multiple Auto Scaling Groups (ASGs). By systematically modifying these variables, the goal was to maintain or improve the performance observed in the baseline configuration while achieving cost efficiency.

PERF-961 - Getting issue details... STATUS  

Summary

  • Through a series of experiments involving different placement strategies, instance types, and total instance counts, we found that the performance remained consistent when using these two configuration: 
    • six c7g.large instances (three dedicated to the okapi service and three allocated to mod-courses, mod-sender, mod-tasks-list, mod-gobi, edge-dematic, mod-erm-usage, mod-eusage-reports, mod-notify, and mod-data-import services) alongside ten r7g.xlarge instances for all other services, with the CPU parameter set to 2 for all services.
    • three c7g.large instances dedicated to the okapi service alongside five r7g.2xlarge instances for all other services, with the CPU parameter set to 2 for all services.
  • Notably, both environment configuration offers a 40% cost reduction compared to the existing setup, making it a more economical option without compromising on performance.
  • Configurations with three c7g.large instances for the okapi service and five r7g.2xlarge instances for all other services show the best performance across all experiments. So it will be more tests in this ticket.
  • Tests had 100% errors count for AIE_TC: Create Invoices, AIE_TC: Invoices Approve, AIE_TC: Paying Invoices, TC: Receiving-an-Order-Line, Unreceiving-a-Piece and Unreceiving-a-Piece Workflows because data was not regenerated.  

Test Runs

Test #DescriptionStatus
Test 1CPU=2 was set for all services, used two autoscaling groups, 1st with 3 Instance Type: c7g.large for okapi service and 10 Instance Type: r7g.xlarge for others services. Completed
Test 2CPU=2 was set for all services, used two autoscaling groups, 1st with 3 Instance Type: c7g.large for okapi service and 10 Instance Type: m6g.2xlarge for others modules. Completed
Test 3CPU=2 was set for all services, used two autoscaling groups, 1st with 3 Instance Type: c7g.large for okapi service and 10 Instance Type: m6g.2xlarge for others services (Repeat Test 2).Completed
Test 4CPU=2 was set for all modules, used two autoscaling groups, 1st with 6 Instance Type: c7g.large, 3 of them for okapi service and other 3 for mod-cources, mod-sender, mod-tasks-list, mod-gobi, edge-dematic, mod-erm-usage, mod-eusage-reports, mod-notify, mod-data-import services and 10 Instance Type: r7g.xlarge for others modules.Completed
Test 5CPU=2 was set for all modules, used two autoscaling groups, 1st with 6 Instance Type: c7g.large, 3 of them for okapi service and other 3 for mod-cources, mod-sender, mod-tasks-list, mod-gobi, edge-dematic, mod-erm-usage, mod-eusage-reports, mod-notify, mod-data-import services and 10 Instance Type: r7g.xlarge for others modules (Repeat Test 4).Completed
Test 6CPU=2 was set for all modules, used two autoscaling groups, 1st with 6 Instance Type: c7g.large, 3 of them for okapi service and other 3 for mod-cources, mod-sender, mod-tasks-list, mod-gobi, edge-dematic, mod-erm-usage, mod-eusage-reports, mod-notify, mod-data-import services and 10 Instance Type: r7g.xlarge for others modules. Test was run without DE_Exporting MARC Bib records custom workflow, DE_Exporting MARC Bib records workflow, and OPIH_/oai/records workflow workflows.Completed
Test 7CPU=2 was set for all modules, used two autoscaling groups, 1st with 6 Instance Type: c7g.large, 3 of them for okapi service and other 3 for mod-cources, mod-sender, mod-tasks-list, mod-gobi, edge-dematic, mod-erm-usage, mod-eusage-reports, mod-notify, mod-data-import services and 10 Instance Type: r7g.xlarge for others modules. Test was run without DE_Exporting MARC Bib records custom workflow, DE_Exporting MARC Bib records workflow, and OPIH_/oai/records workflow workflows (Repeat Test 6).Completed
Test 8CPU=2 was set for all modules, used two autoscaling groups, 1st with 6 Instance Type: c7g.large, 3 of them for okapi service and other 3 for mod-cources, mod-sender, mod-tasks-list, mod-gobi, edge-dematic, mod-erm-usage, mod-eusage-reports, mod-notify, mod-data-import services and 10 Instance Type: r7g.xlarge for others modules. Completed
Test 9CPU=2 was set for all modules, used two autoscaling groups, 1st with 6 Instance Type: c7g.large, 3 of them for okapi service and other 3 for mod-cources, mod-sender, mod-tasks-list, mod-gobi, edge-dematic, mod-erm-usage, mod-eusage-reports, mod-notify, mod-data-import services and 10 Instance Type: r7g.xlarge for others modules. Test was run without DE_Exporting MARC Bib records custom workflow, DE_Exporting MARC Bib records workflow, and OPIH_/oai/records workflow workflows (Repeat Test 7 after Terminate Instances r7g.xlarge).Completed
Test 10CPU=2 was set for all modules, used two autoscaling groups, 1st with 6 Instance Type: c7g.large, 3 of them for okapi service and other 3 for mod-cources, mod-sender, mod-tasks-list, mod-gobi, edge-dematic, mod-erm-usage, mod-eusage-reports, mod-notify, mod-data-import services and 10 Instance Type: r7g.xlarge for others modules (Repeat Test 8 after Terminate Instances r7g.xlarge).Completed
Test 11CPU=2 was set for all services, used two autoscaling groups, 1st with 3 Instance Type: c7g.large for okapi service and 11 Instance Type: r7g.xlarge for others services. Completed
Test 12CPU=2 was set for all services, used two autoscaling groups, 1st with 3 Instance Type: c7g.large for okapi service and 11 Instance Type: r7g.xlarge for others services (Repeat Test 11).Completed
Test 13CPU=2 was set for all services, used two autoscaling groups, 1st with 3 Instance Type: c7g.large for okapi service and 11 Instance Type: r7g.xlarge for others services (Repeat Test 11).Completed
Test 14CPU=2 was set for all services, used two autoscaling groups, 1st with 3 Instance Type: c7g.large for okapi service and 11 Instance Type: r7g.xlarge for others services (Repeat Test 11 after terminate 11 instances).Completed
Test 15CPU=2 was set for all services, used two autoscaling groups, 1st with 3 Instance Type: c7g.large for okapi and mod-authtoken services and 11 Instance Type: r7g.xlarge for others services. Completed
Test 16CPU=2 was set for all services, used two autoscaling groups, 1st with 3 Instance Type: c7g.large for okapi service and Instance Type: r7g.2xlarge for others services. Completed
Test 17CPU=2 was set for all services, used two autoscaling groups, 1st with 3 Instance Type: c7g.large for okapi service and Instance Type: r7g.2xlarge for others services (Repeat Test 16). Completed

Test Results

This table contains durations for all Workflows. 

WorkflowsTest 1

Test 2

Test 3

(repeat 2)

Test 4

Test 5

(repeat 4)

Test 6
 

Test 7

(repeat 6 )

Test 8

Test 9

(repeat 7 after Terminate)

Test 10

(repeat 8 after Terminate)

Test 11

 
Test 12
 
Test 13
 

Test 14

(Terminate)

Test 15Test 16 Test 17 

Average response time (milliseconds)ErrorsAverage response time (milliseconds)ErrorsAverage response time (milliseconds)ErrorsAverage response time (milliseconds)ErrorsAverage response time (milliseconds)ErrorsAverage response time (milliseconds)ErrorsAverage response time (milliseconds)ErrorsAverage response time (milliseconds)ErrorsAverage response time (milliseconds)ErrorsAverage response time (milliseconds)ErrorsAverage response time (milliseconds)ErrorsAverage response time (milliseconds)ErrorsAverage response time (milliseconds)ErrorsAverage response time (milliseconds)ErrorsAverage response time (milliseconds)ErrorsAverage response time (milliseconds)ErrorsAverage response time (milliseconds)Errors
AIE_TC: Create Invoices10227100%7425100%7645100.00%8563100.00%8132100.00%7383100.00%7575100.00%7359100.00%10061100.00%11269100.00%8607100%11352100%8293100%9397100%10687100%8138100%7063100%
AIE_TC: Invoices Approve4649100%3281100%3252100.00%3442100.00%3175100.00%3081100.00%2966100.00%3100100.00%5464100.00%7132100.00%3438100%4792100%3506100%4052100%5297100%3054100%2873100%
AIE_TC: Paying Invoices5686100%3375100%3481100.00%3439100.00%3250100.00%3079100.00%2948100.00%3115100.00%4504100.00%4245100.00%3580100%4635100%3548100%4387100%6265100%3093100%2811100%
CICO_TC_Check-In Controller38280%23500%24460.00%23250.00%22590.00%20920.00%20460.00%21220.00%31130.00%29650.00%25080%31620%23960%28960%38170%21540%19020%
CICO_TC_Check-Out Controller59600%40850%42730.00%41290.00%39050.00%37690.00%35490.00%38160.00%51670.00%52240.00%43070%59541%44060%51000%67880%37210%34360%
CSI_TC:Share local instance1304519%160810%160910.00%1310518.63%1304618.98%160870.00%1311918.66%160710.00%160760.00%160580.00%43070%1326118%160540%1312319%160710%1307619%160050%
DE_Exporting MARC Bib records custom workflow534840%1013240%468600.00%971690.00%1600220.00%



442920.00%

723720.00%495150%17913638%648270%1273730%632060%849000%694460%
DE_Exporting MARC Bib records workflow427370%922130%648970.00%777470.00%2656510.00%



376220.00%

705570.00%456400%18790941%630850%1594530%842200%728050%699050%
EVA_TC: View Account8220%5780%5820.00%5780.00%5530.00%5230.00%5060.00%5340.00%35390.00%27090.00%6020%15132%5270%6700%9670%5030%4800%
ILR_TC: Create ILR19790%16050%16450.00%15950.00%15230.00%14360.00%13940.00%15280.00%31470.00%31220.00%15760%21240%17080%19280%26430%14960%12980%
MSF_TC: mod search by auth query222726%16720%8240.00%9332.15%6860.00%6860.00%6800.00%6920.00%6930.00%6220.00%7700%12240%6800%8100%10790%6660%6690%
MSF_TC: mod search by boolean query4482%4060%2010.00%2570.41%1720.00%1670.00%1670.00%1680.00%1670.00%1560.00%1930%2540%1630%2060%2850%1590%1630%
MSF_TC: mod search by contributors8820%9820%4950.00%5661.25%4070.00%4150.00%4020.00%4060.00%3940.00%3770.00%4410%5100%3990%4800%6050%3930%4060%
MSF_TC: mod search by filter query6492%6030%3440.00%3580.23%2850.00%2860.00%2850.00%2910.00%2740.00%2670.00%3150%3430%2880%3310%4070%2850%2850%
MSF_TC: mod search by keyword query6040%5960%3450.00%3740.41%2900.00%2880.00%2860.00%2920.00%2740.00%2630.00%3160%3430%2830%3290%4020%2840%2870%
MSF_TC: mod search by subject query9301%9320%5150.00%5980.80%4200.00%4300.00%4190.00%4250.00%4080.00%3900.00%4560%5630%4160%4850%5880%4130%4170%
MSF_TC: mod search by title query23810%19650%12600.00%13620.49%10590.00%10610.00%10540.00%10640.00%10260.00%10120.00%10930%11570%10620%11350%12480%10640%10810%
OPIH_/oai/records50860%32480%32540.00%51930.00%30690.00%



41130.00%

24510.00%62670%48850%8820%30260%9430%29800%9400%
POO_TC: Add Order Lines601931%533410%545070.00%550080.00%546460.00%528060.00%521150.00%540830.00%494750.00%446150.00%565550%654362%592070%597320%621160%556610%530520%
POO_TC: Approve Order452940%417500%419350.00%424450.00%424860.00%415690.00%411940.00%420340.00%373100.00%339820.00%442550%461330%461890%452380%473870%425640%415070%
POO_TC Create Order343050%305830%309650.00%318530.00%314400.00%309700.00%303920.00%305750.00%280810.00%254510.00%326010%422671%330580%339960%228280%317450%304790%
RTAC_TC: edge-rtac40500%22280%21840.00%36650.00%37870.00%19780.00%38880.00%19800.00%92970.00%100410.00%38680%83480%21150%38540%22360%37290%20050%
SDIC_Single Record Import (Create)1478319%118820%114410.00%1386818.39%1334518.82%109640.00%1294918.23%110010.00%120150.00%115560.00%1384818%1623618%125380%1448218%337210%1335718%107020%
SDIU_Single Record Import (Update)223474%196850%192980.00%195040.00%37870.00%188690.00%187060.00%187480.00%217450.00%208290.00%197410%233763%207410%209330%228280%194200%183300%
TC: Receiving-an-Order-Line50466100%44954100%46684100.00%46396100.00%45548100.00%45629100.00%44356100.00%45106100.00%40982100.00%36920100.00%47570100%53302100%50146100%49516100%52266100%46404100%44880100%
Serials-Receiving-Workflow51572100%46047100%47460100.00%48142100.00%47661100.00%45645100.00%45404100.00%46299100.00%42105100.00%38241100.00%49262100%61545100%50285100%50943100%52493100%48123100%45827100%
Unreceiving-a-Piece9422100%8028100%8303100.00%8247100.00%7865100.00%7897100.00%7748100.00%7887100.00%7468100.00%6812100.00%8438100%9945100%8544100%8829100%9368100%8058100%7581100%
ULR_TC: Users loan Renewal Transaction40270%31980%32750.00%30600.00%29570.00%28140.00%28100.00%29690.00%446320.00%886020.00%31970%50801%34060%35830%49550%28650%26260%

Comparison

This graph shows the durations of all workflows compared between the best test result from Test №8, MCPT Basile configuration with 14 instances m6g.2xlarge and result from mcpt tunning report.

Average Case workflows: Part 1.

 

Average Case workflows: Part 2.


Average Case workflowsPart 3.


Test №1

Introduction: The Baseline MCPT Environment configuration was applied, and CPU=2 was set for all services, used two autoscaling groups, 1st with 3 Instance Type: c7g.large for okapi service and 10 Instance Type: r7g.xlarge for others services. Fixed Load (average case) MOBIUS test was run.

Objective: The objective of test was to evaluate the performance of the MCPT environment by applying the baseline configuration, while adjusting the CPU allocation for all services to 2 vCPUs. Additionally, the test aimed to assess the impact of using two distinct autoscaling groups with different instance types: c7g.large for the Okapi service and r7g.xlarge for all other services. By running the Fixed Load (average case) MOBIUS test, the goal was to determine if this modified configuration could achieve similar or improved performance compared to the baseline, while potentially optimizing resource allocation and cost.

Results: Almost all workflows shows significant performance degradation.

Instance CPU Utilization

Service CPU Utilization

Here we can see that okapi modules used 42k% CPU power parameter CPU=2 for module.

Service Memory Utilization

Here we can't see any sign of memory leaks on every module. Memory shows stable trend.


Kafka metrics


DB CPU Utilization

DB CPU was 99% average with ERW: Exporting Receiving Information.

DB Connections

Max number of DB connections was 5600 in maximum.

Test №2

Introduction: The Baseline MCPT Environment configuration was applied, and CPU=2 was set for all services, used two autoscaling groups, 1st with 3 Instance Type: c7g.large for okapi service and 10 Instance Type: m6g.2xlarge for others modules. Fixed Load (average case) MOBIUS test was run.

Objective: The objective of this test was to maintain the use of m6g.2xlarge instances for the main modules, as in the Baseline configuration, while introducing a second autoscaling group with c7g.large instances specifically for the Okapi service. This setup aims to explore the performance and cost-effectiveness of separating the Okapi service onto a smaller instance type while keeping the main services on the larger m6g.2xlarge instances.

Results: We observed nearly identical performance results in over half of the workflows compared to the Baseline configuration.

Instance CPU Utilization

Service CPU Utilization

Here we can see that okapi used 47k% of the CPU power parameter CPU=2 for module.

Service Memory Utilization

Here we can't see any sign of memory leaks on every module. Memory shows stable trend.


Kafka metrics

DB CPU Utilization

DB CPU was 99% average with ERW: Exporting Receiving Information

DB Connections

Max number of DB connections was 5710.

Test №3

Introduction: The Baseline MCPT Environment configuration was applied, and CPU=2 was set for all services, used two autoscaling groups, 1st with 3 Instance Type: c7g.large for okapi service and 10 Instance Type: m6g.2xlarge for others services. Fixed Load (average case) MOBIUS test was run (Repeat Test 2).

Objective: The objective of this test was to validate the consistency of performance observed in Test 2 by repeating the same configuration. 

ResultsWe observed nearly identical performance results almost for all the workflows compared to the Baseline configuration.

Instance CPU Utilization

Service CPU Utilization

Here we can see that okapi used 45k% of the CPU power of the parameter CPU=2.

Service Memory Utilization

Here we can't see any sign of memory leaks on every module. Memory shows stable trend.


Kafka metrics


DB CPU Utilization

DB CPU was 99% maximum.

DB Connections

Max number of DB connections was 5500.

Test №4

Introduction:  The Baseline MCPT Environment configuration was applied, and CPU=2 was set for all modules, used two autoscaling groups, 1st with 6 Instance Type: c7g.large, 3 of them for okapi service and other 3 for mod-cources, mod-sender, mod-tasks-list, mod-gobi, edge-dematic, mod-erm-usage, mod-eusage-reports, mod-notify, mod-data-import services and 10 Instance Type: r7g.xlarge for others modules. Fixed Load (average case) MOBIUS test was run.

Objective: The objective of this test was to add more services with lower activity levels to the first autoscaling group, increasing the number of c7g.large instances from 3 to 6. For the main services, r7g.xlarge instances were used with a total of 10 instances, which are more cost-effective compared to the Baseline configuration with r6g.4xlarge.

Results: We observed a better performance compared to previous test.

Instance CPU Utilization

Service CPU Utilization

Here we can see that okapi used 43k% of the CPU power of parameter CPU=2.

Service Memory Utilization

Here we can't see any sign of memory leaks on every module. Memory shows stable trend.

Kafka metrics



DB CPU Utilization

DB CPU was 99%.

DB Connections

Max number of DB connections was 5810.

DB load

                                                                                                                    

Top SQL-queries


Test №5

Introduction: The Baseline MCPT Environment configuration was applied, and CPU=2 was set for all modules, used two autoscaling groups, 1st with 6 Instance Type: c7g.large, 3 of them for okapi service and other 3 for mod-cources, mod-sender, mod-tasks-list, mod-gobi, edge-dematic, mod-erm-usage, mod-eusage-reports, mod-notify, mod-data-import services and 10 Instance Type: r7g.xlarge for others modules. Fixed Load (average case) MOBIUS test was run (Repeat Test 4).

Objective: The objective of this test was to validate the consistency of performance observed in Test 4 by repeating the same configuration.

Results: We confirmed performance improvements with this configuration, also a several workflows shoes better performance than in Test 4.


Instance CPU Utilization


Service CPU Utilization

Here we can see that okapi used 46k% of the CPU power of parameter CPU=2.

Service Memory Utilization

Here we can't see any sign of memory leaks on every module. Memory shows stable trend.


Kafka metrics


DB CPU Utilization

DB CPU was 99%.

DB Connections

Max number of DB connections was 5500.

DB load

Top SQL-queries

Test №6

Introduction: The Baseline MCPT Environment configuration was applied, and CPU=2 was set for all modules, used two autoscaling groups, 1st with 6 Instance Type: c7g.large, 3 of them for okapi service and other 3 for mod-cources, mod-sender, mod-tasks-list, mod-gobi, edge-dematic, mod-erm-usage, mod-eusage-reports, mod-notify, mod-data-import services and 10 Instance Type: r7g.xlarge for others modules. Fixed Load (average case without DE_Exporting MARC Bib records custom workflow, DE_Exporting MARC Bib records workflow, and OPIH_/oai/records workflow) MOBIUS test was run.

Objective: The goal of this test was to exclude several unstable workflows and observe whether their removal would have a noticeable impact on the performance of the remaining workflows. Specifically, by omitting the DE_Exporting MARC Bib records custom workflow, DE_Exporting MARC Bib records workflow, and OPIH_/oai/records workflow, the test aimed to assess if the stability and performance of the other workflows would improve under the same configuration.

Results: We observed minor performance improvements across almost all workflows. However, these improvements were not substantial. It appears that the observed effects are relatively small and can be attributed to a reduced overall load in the environment.

Instance CPU Utilization

Service CPU Utilization

Here we can see that okapi used 46k% of the CPU power of parameter CPU=2.

Service Memory Utilization

Here we can't see any sign of memory leaks on every module. Memory shows stable trend.

Kafka metrics


DB CPU Utilization

DB CPU was 99%.

DB Connections

Max number of DB connections was 6000.

DB load

Top SQL-queries


Test №7 - 8

Introduction:  

  • Test 7: The Baseline MCPT Environment configuration was applied, and CPU=2 was set for all modules, used two autoscaling groups, 1st with 6 Instance Type: c7g.large, 3 of them for okapi service and other 3 for mod-cources, mod-sender, mod-tasks-list, mod-gobi, edge-dematic, mod-erm-usage, mod-eusage-reports, mod-notify, mod-data-import services and 10 Instance Type: r7g.xlarge for others modules. Fixed Load (average case without DE_Exporting MARC Bib records custom workflow, DE_Exporting MARC Bib records workflow, and OPIH_/oai/records workflow) MOBIUS test was run (Repeat Test 6).
  • Test 8: The Baseline MCPT Environment configuration was applied, and CPU=2 was set for all modules, used two autoscaling groups, 1st with 6 Instance Type: c7g.large, 3 of them for okapi service and other 3 for mod-cources, mod-sender, mod-tasks-list, mod-gobi, edge-dematic, mod-erm-usage, mod-eusage-reports, mod-notify, mod-data-import services and 10 Instance Type: r7g.xlarge for others modules. Fixed Load (average case) MOBIUS test was run.

Objective: The objective of these tests was to validate the consistency of performance observed in Test 5 and Test 6. This was achieved by repeating the same configuration.

Results: We see performance improvements for Test 8.

Instance CPU Utilization

Service CPU Utilization

Here we can see that okapi used 48k% of the CPU power of parameter CPU=2.

Service Memory Utilization

Here we can't see any sign of memory leaks on every module. Memory shows stable trend.