Host ai workflow runner with Spring Boot: streamed responses

Summary

Deploy an AI workflow runner built with Spring Boot on Ample using the streaming response service pattern. Compute runs the app in an isolated microVM behind a public HTTPS URL, with a managed PostgreSQL 16 database auto-provisioned and injected as DATABASE_URL. Verified on Spring Boot: a Server-Sent Events endpoint read incrementally through the public HTTPS gateway, with five one-second chunks arrived spread over time. This confirmed a 70-second stream of 36 chunks, which surpassed common 60-second idle timeouts and observed a client disconnect after two chunks (disconnected=true).

Infrastructure requirements

• Compute: verified (Apps run in isolated x86_64 Firecracker microVMs that auto-pause when idle and wake on request; sizes are the priced VM sizes.)
• Postgres: verified (Managed PostgreSQL 16 runs in its own microVM and is auto-provisioned when an app needs a database and no DATABASE_URL is supplied.)

Prerequisites

• A Spring Boot project that builds and starts with documented commands (./mvnw -q -DskipTests package or the Gradle wrapper).
• A PostgreSQL driver reading DATABASE_URL at runtime (auto-provisioned when omitted, or supplied with --env).
• An Ample account token with servers:write, databases:read.

Exact tested configuration

• template: jvm-21
• runtime: java
• size: s-1vcpu-2gb
• build: ./mvnw -q -DskipTests package
• start: java -jar $(ls target/*.jar | grep -v -- '-plain' | head -n 1)

Steps

1. Build and start. Run ./mvnw -q -DskipTests package (or the Gradle wrapper) to produce one application jar, then java -jar on the jvm-21 template (Temurin JDK 21) with server.port read from PORT. The server must bind 0.0.0.0 on PORT.
2. Implement the pattern on PostgreSQL. The fixture's module implements streaming response service: a Server-Sent Events endpoint read incrementally through the public HTTPS gateway, conducting a 70-second stream of 36 chunks. Follow the approach in your schema; ensure migrations are idempotent and run them with --release-command.
3. Deploy. Run the synchronous deploy once to read the result:

   ample deploy . --name <app-name> --public
   

4. Verify. Fetch the live URL and self-test route(s) from the example; run own checks. On failure read ample logs <deployment_id> --kind build then --kind runtime.

   ample logs <deployment_id> --kind build
   

Success checks

• App responds on its public URL (/ on the live URL).
• Streaming-response-service self-test (/p/streaming-response-service on the live URL).

Limitations

• Verified on the jvm-21 template at s-1vcpu-2gb; other sizes and configurations are not verified.
• Apps auto-pause when idle and wake on the next request; always-on is an operator setting.

Cost estimate

Estimated 10.00 USD per month:

• app server x1 s-1vcpu-1gb: 5.00 USD
• managed PostgreSQL database x1 s-1vcpu-1gb: 5.00 USD

Verification evidence

• Successful canary run at revision showing the Spring Boot pattern fixture deployed on Ample.

Execution binding

MCP tool ample_deploy, schema hash 876465fce906da0c, binding state current, required scopes: servers:write, databases:read.