⏺ Shift-Left API Development: Automated Testing & CI/CD with the Apigee Emulator | Community TechTalk recording

Hello Apigee Community! :waving_hand:

In our latest TechTalk, Google Cloud Apigee Specialist Customer Engineer Nigel Walters demonstrates how to implement shift-left testing workflows for Apigee API proxies. Nigel explains how to use the containerized Apigee Local Emulator and the open-source apigee-emulator-service to bypass slow remote cloud deployment cycles and validate proxy logic in milliseconds.

:television: Watch the recording

:file_folder: Access the presentation slides

Shift-left testing and the enterprise QA pyramid :gear:

Waiting on remote control plane deployments during early development introduces friction and slows release cycles. Nigel outlined three core layers to balance testing speed, cost, and coverage:

  • Linting and static analysis: Catch syntax errors, anti-patterns, and naming violations early using apigeelint alongside language-specific linters for JavaScript, Java, or Python policies.

  • Local emulator validation: Execute sub-second assertion suites against a containerized message processor with embedded Cassandra persistence, validating Key Value Maps, target servers, API products, and developer apps locally.

  • Cloud runtime verification: Reserve remote Apigee X and Apigee Hybrid environments for end-to-end integration testing, real identity provider flows, and live backend connections.

Key takeaways and demonstration highlights :light_bulb:

Nigel walked through a hands-on demonstration using Google Cloud Shell:

  • Declarative deployment manifests: Defined proxy endpoints, target servers, policies, developer apps, and test assertions within a single YAML file powered by Apigee Feature Templator (AFT) syntax.

  • Automated assertion execution: Demonstrated automated deployment and testing via local shell scripts, filtering test execution outputs and status assertions with jq.

  • Execution tracing and web UI: Visualized request flow, variable evaluation, and policy execution steps directly within the emulator service’s web console.

  • Shared preview sandboxes on Cloud Run: Deployed the emulator service container to Google Cloud Run, exposing an authenticated public endpoint for remote QA teams to test without needing dedicated cloud environments.

Resources :books:

Next steps :chequered_flag:

  • :thought_balloon: Continue the conversation: How is your team managing local proxy unit testing and CI/CD quality gates? Share your experiences or ask Nigel questions in the comments below!

  • :star: Share your feedback: Help us improve future sessions by completing our brief feedback poll!

  • :megaphone: Stay tuned: Register for our upcoming Community TechTalks, including our session on modern API management to AI gateway with Geir Sjurseth on October 22, and end-to-end AI workflows with Luis on October 29 → Google Cloud Apigee Community TechTalks

3 Likes

The shift-left approach makes a lot of sense for API proxy development. Being able to run fast local assertions and CI checks before deploying to Apigee should make it much easier to catch issues early and keep cloud environments focused on integration and end-to-end validation. The Cloud Run preview environment also sounds useful for teams that need a shared testing endpoint.

for local emulator setup i am getting wait time exceed for all the docker emulator version , i hope i will get proper resolution to rectify this

The shift-left approach looks really useful, especially for reducing the feedback loop during API development. I’m particularly interested in how the local emulator compares with live Apigee environments when testing authentication, external API dependencies, and error-handling scenarios.

The CI/CD integration also seems valuable for catching proxy issues before deployment. Looking forward to seeing the practical examples and understanding how teams decide which tests should stay local versus being validated in the cloud.

1 Like

This TechTalk hits on something teams often skip validating proxy logic locally before touching a cloud environment.

The three-layer pyramid (lint → local emulator → cloud runtime) is the right mental model. In practice, the hardest part is getting teams to trust the emulator enough to stop defaulting to remote deployments for every small change.

One thing I’m curious about: how does the emulator handle Apigee Hybrid-specific behaviors particularly around Istio sidecar traffic and custom target server configurations? That’s usually where local validation breaks down for on-prem deployments.

1 Like

Yes, indeed, the key point is trust. You have to know that your tests are valid for them to have value - we ensure this from the Apigee side by shipping the same runtime inside the emulator as we provide for Apigee X / Hybrid.

Target server configurations are supported in the emulator (https://docs.cloud.google.com/apigee/docs/api-platform/local-development/vscode/deploy-environment#target-servers) though if you are using the deployment.yaml for the emulator service ( GitHub - tyayers/apigee-emulator-service: A service to run the Apigee emulator in a container with a UI frontend. · GitHub ) then you’ll have to specify the target configuration in the proxy yaml configuration.

If in relation to Istio sidecar traffic you are referring to the Apigee Envoy Adapter then it would be possible to also deploy that locally as part of the local test setup.

1 Like

Thanks Nigel really appreciate you taking the time to answer this.

The fact that the same runtime ships inside the emulator changes everything. That’s the argument I needed to convince teams to stop treating remote deployments as the default for every small change.

Will definitely explore running the Apigee Envoy Adapter locally as part of the test setup that’s a practical path I hadn’t considered for Hybrid environments.

See you on October 22

1 Like