Goal

The goal of this epic is to facilitate performance testing of the ACM and CNV integration, particularly concerning the search API and its ability to manage a high volume of VMs. This will ensure the system can support large customer environments effectively, identifying and addressing backend performance bottlenecks before they impact UI performance or customer experience.

User Stories

• As a performance engineer, I want to stress-test the ACM search API with a large number of VMs (e.g., 100,000 across 20 clusters) so that we can identify and resolve performance bottlenecks in the PostgreSQL database and API before customer deployment.
• As a cluster administrator, I want to have a robust testing environment with a high volume of VMs readily available so that I can accurately assess the performance of ACM and CNV integration features, including VM operations like creation and deletion.
• As an OpenShift developer, I want a clear communication channel with the ACM search API team and access to deployment automation so that we can efficiently set up test environments and collaborate on resolving performance issues.

Non-Requirements

• This epic does not include the development of new UI features for ACM 2.15.
• This epic does not cover performance testing of ACM features unrelated to CNV integration or VM management.
• Direct UI performance optimization is not the initial focus; backend performance (especially the search API and PostgreSQL) will be addressed first.
• This epic does not include the implementation of pagination for the search API, as this is slated for release 2.16.

Notes

• Collaboration with Alex Kros from the ACM Perf&Scale department is crucial for deployment automation and tooling.
• API documentation is necessary to facilitate stress-testing of PostgreSQL tables.
• The current limitation of the search API, lacking pagination, means initial large-scale testing will focus on backend data pushing and querying rather than full end-to-end UI scenarios involving 100,000 VMs.
• Communication with developers from the search API team, will be essential for technical guidance and issue resolution.
• The target for VM capacity is 100,000 VMs across 20 clusters (5,000 VMs per cluster).
• We will begin the VM deployment on the ACM multicluster solution in the following manner to simplify the performance testing process:
  • ACM with 10K VMs across multiple clusters.
  • ACM with 20K VMs across multiple clusters.
  • ACM with 50K VMs across multiple clusters.
  • ACM with 100K VMs across multiple clusters.