
kgateway
The Cloud-Native API Gateway and AI Gateway
- 8
- Projects
- 2
- Years active
- 1
- In 2026 T2
Projects
82026
4 projectsTerm 2 · Jun-Aug · 1 project
Documentation improvements for OIDC and JWT IdPs integrations
LFX Mentorship · Listed in the official cohort
The goal of this project is to test and document common JWT validation scenarios (e.g., JWKS, multiple issuers, and claim-based validation) and demonstrate kgateway integration with multiple OAuth identity providers.
- Technical writing
- Kubernetes
- Understanding of authentication/authorization concepts (JWT
- OAuth2
- etc.)
- Git
Term 1 · Mar-May · 3 projects
Add support for Chaos Engineering/Fault Injection
LFX Mentorship · Listed in the official cohort
This project focuses on adding fault injection support in kgateway, enabling platform operators and developers to test system resiliency under controlled failure scenarios. Fault injection allows teams to proactively identify weaknesses by introducing network latency, service errors, or resource constraints. This project will involve designing a configuration API for specifying fault injection rules, implementing support in the kgateway plugin framework, integrating with Envoy’s native fault injection capabilities, and creating documentation and examples to demonstrate practical use cases in Kubernetes environments.
- Go
- Kubernetes
- Kubernetes Gateway API
- Envoy
Ecosystem Integrations & Tutorials for kgateway's AI Gateway
LFX Mentorship · Listed in the official cohort
This project focuses on creating clear, approachable, and practical documentation that shows how to integrate kgateway’s AI Gateway (agentgateway) with popular AI tools, developer UIs, and other CNCF ecosystem projects. Its aim is to make kgateway’s documentation more practical, discoverable, and reflective of real ecosystem usage, so users can better understand what’s possible and how kgateway fits into the broader AI and cloud-native landscape. Examples of integrations and tutorials include:
- Strong written communication skills
- Interest in learning and exploring new projects!
- Basic understanding of GitHub
- Markdown
- and technical blogging
- (Bonus) Experience using any of the following: Open WebUI
- +4
Modernize and improve kgateway end-to-end testing
LFX Mentorship · Listed in the official cohort
Kgateway has a large number of end-to-end (e2e) tests that functionally test the full end-to-end flow of various routing and policy features present for both supported dataplane proxies, Envoy and agentgateway. However, these tests have been implemented with a custom e2e framework which is starting to show some cracks, notably around the speed of execution and general obtuseness of the framework. This project will contain two main phases: 1. Analyze the gaps of the current custom e2e framework and work with kgateway maintainers to decide whether or not it is worth migrating to an alternative solution 2. Depending on the outcome of phase 1, identify a set of e2e tests to either migrate to the new framework or improve the execution of in the current framework This project will provide exposure to a wide range of features in kgateway as well as real-world experience with ensuring the kgateway project is providing rock solid reliability for its users!
- Go
- Kubernetes
- Kubernetes Gateway API
- End-to-end testing experience
2025
4 projectsTerm 3 · Sep-Nov · 2 projects
Improve kgateway ecosystem integrations documentation
LFX Mentorship · Listed in the official cohort
This project aims to create clear, approachable documentation and blog posts that showcase how to integrate kgateway with key CNCF projects such as Argo, Istio, and KServe. These resources will make it easier for users to adopt kgateway in modern cloud-native environments and promote deeper integration across the CNCF ecosystem.
- Strong written communication skills
- Interest in learning and exploring new projects!
- Basic understanding of GitHub
- Markdown
- and technical blogging
- (Bonus) Experience using any of the following: Argo
- +2
kgateway, agentgateway, and observability improvements
LFX Mentorship · Listed in the official cohort
This project focuses on enhancing observability in kgateway’s agentgateway integration by adding support for OpenTelemetry-based tracing, exposing Prometheus-compatible metrics, and configuring access logging. It builds on a previous LFX project that introduced tracing support for kgateway’s AI extproc server, extending those capabilities to support span propagation from agentgateway. In addition to adding support for configuring the kgateway control plane to set agentgateway’s tracing configuration, we want to test exporting traces to OpenTelemetry-compatible backends such as the OpenTelemetry Collector and Jaeger, and create user-facing documentation for these integrations. Beyond emitting basic spans, this project will also document how to enrich traces with authenticated user context derived from RBAC decisions, enabling more fine-grained debugging and attribution.
- Go
- Rust
- Otel
- Kubernetes
- Kubernetes Gateway API
- (Bonus) Familiarity with MCP and/or A2A
Term 2 · Jun-Aug · 2 projects
Mission: Scale Possible! Build Automated Scale Tests for kgateway
LFX Mentorship · Listed in the official cohort
This project aims to design and build a suite of automated scale tests for kgateway that would run as part of our builds and releases process. These tests will help us understand how kgateway behaves under a large number of Kubernetes gateway resources and kgateway extensions, and load tests to ensure it performs reliably as usage scales. You'll gain hands-on experience with performance testing, infrastructure automation, and Kubernetes-based systems.
- Go
- GitHub workflow
- Kubernetes
OpenTelemetry is an AI Gateway’s Best Friend: Extending Observability to kgateway’s AI Extensions
LFX Mentorship · Listed in the official cohort
Kgateway can be used as an “AI Gateway” that allows applying traditional traffic management, security, and resiliency policies when reaching out to LLM providers. It also allows sending a request to another server for processing things like prompt guards or enrichment before we send it off to the LLM providers. An important aspect in any system that has lots of moving parts is observability: the collection of metrics and other information that allow you to identify issues and troubleshoot a live system. This project aims to enable OpenTelemetry support in kgateway’s AI extensions by allowing users to configure a gRPC tracing collector. This feature enables span propagation from the AI extensions extproc server and exporting traces to an OpenTelemetry-compatible backend (e.g., OpenTelemetry Collector, Jaeger, Zipkin, or Datadog). This feature will build on Envoy’s native tracing capabilities and aligns with industry-standard observability practices. This will make it easier to debug, monitor, and optimize traffic flowing to LLM Providers through kgateway.
- Go
- Python
- GitHub workflow
- Otel
- Kubernetes
- Kubernetes Gateway API
- +1