DocumentationAEM GuidesAEM Guides Documentation

AEM Guides Publishing Benchmarks on AEMaaCS

Last update: May 6, 2026

CREATED FOR:

  • User
  • Admin

This benchmark evaluates the performance of the new publishing APIs across different output presets and increasing map sizes in AEM Guides as a Cloud Service. The goal is to understand scalability behavior and identify performance bottlenecks.

The publishing service uses a microservice-based architecture with autoscaling, enabling handling of larger workloads through additional pods.

Execution environment

  • AEM release:2026.4.25322.20260407T085152Z
  • Guides Add-on release: 2026.5.0
  • Initial pod count: 2 pods
  • Autoscaling behavior: Scaled up to 4 pods across 4 nodes as load increased
  • vCPUs: 10
  • RAM per pod: 8 GB
  • Concurrent users: 1 user
NOTE
This exercise focused on how publishing behaves as map size increases, highlighting the impact of larger maps on throughput, latency, and overall publish completion under load.

Output generation numbers

Native AEM Site

MapSize
Execution Time (s)
Microservice
10
62.378
Yes
100
64.27
Yes
1000
93.091
Yes
5000
496.319
Yes
10000
922.602
Yes

Native PDF

MapSize
Execution Time (s)
Microservice
10
62.302
Yes
100
62.431
Yes
1000
108.666
Yes
5000
201.379
Yes
10000
1170.689
Yes

PDF

MapSize
Execution Time (s)
Microservice
10
30.926
Yes
100
30.987
Yes
1000
77.007
Yes
5000
247.505
Yes
10000
686.61
Yes

HTML5

MapSize
Execution Time (s)
Microservice
10
30.844
Yes
100
30.834
Yes
1000
77.384
Yes
5000
170.237
Yes
10000
419.166
Yes

Key observations

  • AEM Site generation time depends on the template being used. Execution time may increase if a complex template is used.
  • Custom publishing execution time is based on a sample custom output. Custom publishing time solely depends on customer’s own transformation logic.
recommendation-more-help
experience-manager-guides-help-product-guide