Bug 2082084

Summary: Benchmark Octavia dataplane performance
Product: Red Hat OpenStack Reporter: Nate Johnston <njohnston>
Component: openstack-octaviaAssignee: Gregory Thiemonge <gthiemon>
Status: NEW --- QA Contact: Bruna Bonguardo <bbonguar>
Severity: high Docs Contact:
Priority: high    
Version: 17.0 (Wallaby)CC: hrakotor, lpeer, majopela, njohnston, scohen
Target Milestone: ---Keywords: FutureFeature, Tracking
Target Release: ---Flags: sokhan: needinfo+
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:

Description Nate Johnston 2022-05-05 11:39:39 UTC 
Users of Octavia have often requested metrics on the dataplane performance for Octavia load balancers.  Metrics on how many requests per second or megabits per second can be pushed through a single load balancer instance are critical for someone planning the architecture of an application that has Octavia as a key component.  Create a report, consumable by customers, that provides the kind of dataplane performance metrics that customers need.  This should use best-of-breed HTTP load generation tools.

Scenarios that should be measured:

- OSP 17.0 vs. OSP 16.x
- Amphora loadbalancers vs. OVN loadbalancers

Note that control plane performance is not being tested - it is not relevant how many load balancers can be spawned or how many members they can have for the purposes of this study.