Server Configuration

Server Configuration

Everything below is real configuration from evento-server/src/main/resources/application.properties and BusProperties. Spring's relaxed binding applies, so evento.server.bus.port can equally be set as the environment variable evento_server_bus_port.

---

1. Ports

Port	Property	Purpose
3000	server.port	HTTP — GUI and REST API
3030	evento.server.bus.port	Netty message bus — where bundles connect

---

2. Message bus — evento.server.bus.*

Bound by the BusProperties record (@ConfigurationProperties(prefix = "evento.server.bus")). The bus is opt-in: evento.server.bus.enabled=true (the shipped application.properties sets it).

Property	Default	Notes
enabled	(set to true in the shipped config)	Master switch
port	3030	Bus listen port
server-instance-id	server-<random UUID>	Blank/absent generates one
correlation-check-interval	1s	Expiry sweep for pending correlations
shutdown-deadline	30s	Bounded graceful stop
heartbeat-write-idle	15s	Send a Ping after this much write idle
heartbeat-read-idle	45s	Close the channel after this much read idle
connect-timeout	5s
max-frame-length	16 MB	Per-chunk memory, not a message size limit — see Wire Protocol § 3
write-buffer-high-water-mark	1 MB	Channel transitions to DEGRADED above this
write-buffer-low-water-mark	256 KB	Must be strictly less than the high mark, or startup throws
business-executor-core-size	max(8, cores × 2)
business-executor-max-size	cores × 8	Clamped up to at least the core size
business-executor-queue-capacity	256	Do not enlarge to absorb load — see Throughput and Capacity
business-executor-keep-alive	90s
business-executor-saturation-warn-interval	10s	Rate-limits the saturation WARN line
auth-token	(empty)	Empty accepts all bundles; set a shared secret to require it

---

3. Event store — evento.es.*

Property	Default	Purpose
evento.es.mode	APES in code, CPES in the shipped properties	Event store mode
evento.es.fetch.concurrency	4	Fair semaphore in EventStoreRequestHandlerBridge
evento.es.fetch.delay	69	ms between fetch cycles
evento.es.aggregate.state.cache.size	1024	Snapshot cache entries
evento.es.aggregate.state.cache.expiry	150000	ms
evento.es.aggregate.events.cache.size	1024	Event cache entries
evento.es.aggregate.events.cache.expiry	150000	ms

evento.es.fetch.concurrency is very likely the binding constraint on a consumer-heavy cluster. It is a fair semaphore, so no amount of bus threads raises it. It is deliberately left at 4: it guards heap against concurrent EventFetchRequest result sets, and the OOM it was added for is real. Raise it against measured headroom, not on principle.

---

4. Database

spring.datasource.url=${SPRING_DATASOURCE_URL:jdbc:postgresql://localhost:5432/evento}
spring.datasource.username=${SPRING_DATASOURCE_USERNAME:postgres}
spring.datasource.password=${SPRING_DATASOURCE_PASSWORD:secret}
spring.jpa.database-platform=org.hibernate.dialect.PostgreSQLDialect
spring.jpa.hibernate.ddl-auto=none
spring.sql.init.mode=always

Hikari keepalive/timeout settings ship as spring.datasource.hikari.data-source-properties.{tcpKeepAlive,loginTimeout,connectTimeout,socketTimeout}.

---

5. Authentication

Two independent surfaces:

# GUI + REST (HTTP Basic)
spring.security.user.name=${SPRING_SECURITY_USER_NAME:evento}
spring.security.user.password=${SPRING_SECURITY_USER_PASSWORD:secret}
spring.security.user.roles=WEB,ADMIN

# Wire-level bundle auth — blank accepts all bundles (dev only)
evento.server.bus.auth-token=${EVENTO_SERVER_BUS_AUTH_TOKEN:}

Both default to development-grade values. secret as a password and a blank bundle auth token are fine on a laptop and unacceptable anywhere else. See Security Model.

---

6. Telemetry and other settings

Property	Shipped value	Purpose
evento.cluster.name	evento-server	Cluster display name
evento.performance.capture.rate	1 (compose file uses 0.1)	Fraction of invocations sampled
evento.telemetry.ttl	365	Retention, days
spring.servlet.multipart.max-file-size	200MB
spring.mvc.async.request-timeout	3000000

Actuator exposure:

management.endpoints.web.exposure.include=health,info,prometheus,metrics
management.endpoint.health.probes.enabled=true
management.endpoint.health.show-details=when_authorized
management.health.livenessstate.enabled=true
management.health.readinessstate.enabled=true

---

7. Docker image

Published as eventoframework/evento-server on Docker Hub. The image is a two-stage build on digest-pinned eclipse-temurin:25 (JDK to build, JRE to run), exposing 3000 and 3030:

FROM eclipse-temurin:25-jdk-noble@sha256:29d2d8af… AS build
WORKDIR /evento
COPY . .
RUN chmod +x gradlew && ./gradlew :evento-server:bootJar --no-daemon

FROM eclipse-temurin:25-jre-noble@sha256:b27ca476…
EXPOSE 3000
EXPOSE 3030
RUN mkdir /app
COPY --from=build /evento/evento-server/build/libs/*.jar /app/spring-boot-application.jar
ENTRYPOINT ["java","-jar","/app/spring-boot-application.jar"]

Compose files in the repository:

Path	What it runs
docker/evento-docker-compose/docker-compose.yaml	Server + Postgres — the normal way to run Evento
docker/evento-local/docker-compose.yaml	Postgres only, for running the server from an IDE

---

8. Deployment shape

The framework does not manage instance lifecycle. There is no deploy-by-upload, no /spawn or /kill, and no autoscaling protocol — all of that was removed in 2.0. The server emits performance metrics; your orchestrator (k8s, Nomad) owns scaling and restarts. See Migrating from v1.

---

See also

• Throughput and Capacity — how to choose the executor numbers above
• Security Model — hardening the two auth surfaces
• Observability — what to scrape once it is running