Troubleshooting increased deployment times in Adobe Experience Manager as a Cloud Service production pipelines

Last update: October 8, 2026

Learn how to investigate and address increasing deployment durations in AEM as a Cloud Service production pipelines.

Description description

Environment

Adobe Experience Manager as a Cloud Service (AEMaaCS)
Adobe Cloud Manager

Symptoms

Deployments through Adobe Cloud Manager production pipelines take longer to complete than in previous releases.
This issue occurs during the deploy step, specifically when rolling updates are performed on Publish instances. The deployment pipeline typically includes build, unit tests, code quality analysis, and deployment stages.
A significant increase in the time taken for Publish instance start-up and readiness during rolling replacement leads to longer overall deployment durations.
No deployment errors, resource bottlenecks, or failed health checks may be visible in telemetry, and the deployment may still complete successfully.

Cause

The main contributor to increased deployment time is often the serialized rolling replacement of Publish instances, where each new instance must fully start and pass health checks before the next is replaced. Growth in application complexity, repository size, or activation-time processing lengthens this process.

Resolution resolution

Follow the steps below to resolve the issue:

Review recent deployment pipeline executions in Adobe Cloud Manager and compare the duration of each stage, focusing on the deploy step and the rolling update of Publish instances.
Investigate whether recent releases have introduced additional activation-time processing, start-up listeners, or increased the number of OSGi bundles on Publish instances.
Check if the repository size, segment volume, or index count has grown significantly between releases, as these factors can impact Publish instance start-up and readiness times.
During the next production deployment, capture the Publish instance start-up logs to analyze what operations are contributing to the delay. Ensure log capture covers the entire start-up sequence.
If start-up time continues to increase, consider optimizing activation-time processing, reducing unnecessary start-up listeners, or cleaning up unused OSGi bundles and indexes.
After making changes, verify by running another deployment and comparing the deploy step duration to previous runs.

recommendation-more-help

experience-cloud-kcs-help-kbarticles