Overview

Performance Testing Report Overview

This document presents the results of volume testing for Holding records with different item counts.

It includes the results of the impact on other LoC workflows response times when editing holdings records with 10K items attaching to them.

Ticket: https://folio-org.atlassian.net/browse/PERF-1289

Summary

• Volume tests show that Holdings editing response time grows significantly after 1200 items (4,7 sec). 2000 items took almost 13 seconds, 10000 items ~ 1 minute.

• Test editing a single Holdings record while all master script flows running in the background:

  • No observable LoC workflows response times degradation while editing of a Holdings record with 10k items

  • Editing holdings records with 10k under average LoC load increases the editing response time by 30%, rising from 60 seconds to 78 seconds.

  • No crashes or module restarts during tests observed.

  • CPU utilized in accordance with load.

  • No memory issues happened.

• Test editing a Holdings record continuously while all other LoC flows are running in the background (not a realistic scenario)

  • Result: measurable degradation in 218 out of 365 transactions (59.7%). The impact is concentrated in the Finance/Ledger and ListApp operations because of contention on shared backend services like mod-inventory-storage.

Test runs

Volume testing experiments (only tested editing individual holdings records with various number of items attaching to them):

Tests that measured performance of editing the holding records with LoC workfows running in the background:

Results

Response times

Volume testing experiments:

The graph added just to visualize the response times results from T1 - T10, x-axis represents the number of items attached to a holdings record while the y-axis shows the response time in seconds.

Open

Comparison tables