-
Notifications
You must be signed in to change notification settings - Fork 297
Open
Labels
area/test-and-releaseenhancementNew feature or requestNew feature or requesthelp wantedExtra attention is neededExtra attention is neededpriority/P2Nice-to-Have / ExploratoryNice-to-Have / Exploratory
Description
Description
Add a new dynamo profile to the E2E testing framework to test Semantic Router integration with Nvidia Dynamo.
Background
The E2E testing framework introduced in #655 provides an extensible profile-based architecture. We need to add a dynamo profile to test Semantic Router deployment and functionality with Nvidia Dynamo optimization platform.
Tasks
- Create
e2e/profiles/dynamo/directory structure - Implement
Profileinterface for dynamo- Setup: Deploy Nvidia Dynamo components
- Setup: Deploy Semantic Router with Dynamo integration
- Setup: Configure Dynamo optimization settings
- Teardown: Clean up Dynamo resources
- Implement test cases:
- Basic health check with Dynamo
- Optimized inference verification
- Performance comparison (with/without Dynamo)
- Dynamic batching testing
- GPU utilization monitoring
- Add documentation for dynamo profile usage
- Update CI workflow to run dynamo tests
Implementation Details
Profile Structure
type Profile struct {
verbose bool
}
func (p *Profile) Setup(ctx context.Context, opts *framework.SetupOptions) error {
// 1. Deploy Nvidia Dynamo runtime
// 2. Deploy Semantic Router with Dynamo backend
// 3. Configure optimization settings (dynamic batching, kernel fusion)
// 4. Configure GPU resource allocation
}Test Cases
- Dynamo Health Check: Verify Dynamo runtime is healthy
- Optimized Inference: Test inference with Dynamo optimizations
- Performance Benchmark: Compare performance with/without Dynamo
- Dynamic Batching: Verify dynamic batching functionality
- GPU Utilization: Monitor GPU usage and efficiency
Acceptance Criteria
- Dynamo profile can be run with
make e2e-test PROFILE=dynamo - All test cases pass successfully
- Documentation is complete and clear
- CI integration works correctly
- Performance improvements are documented
References
- E2E Framework PR: [Feat] Add automate e2e test framework for extensible integration tests #655
- Nvidia Dynamo Documentation: (add link)
- E2E Framework README:
e2e/README.md
Related Issues
Part of the E2E testing framework expansion effort.
- Related to [E2E] Add Istio profile for E2E testing framework #656 (Istio profile)
- Related to [E2E] Add production-stack profile for E2E testing framework #657 (production-stack profile)
- Related to [E2E] Add llm-d profile for E2E testing framework #658 (llm-d profile)
Metadata
Metadata
Assignees
Labels
area/test-and-releaseenhancementNew feature or requestNew feature or requesthelp wantedExtra attention is neededExtra attention is neededpriority/P2Nice-to-Have / ExploratoryNice-to-Have / Exploratory