Spring boot admin clustering with hazelcast

Spring Boot Admin Server information

• Version:
  3.1.4

• Spring Boot version:
  3.0.9

• Configured Security:
  N/A

• Webflux or Servlet application:
  servlet

Client information

• Spring Boot versions:
  3.0.9

• Used discovery mechanism:
  self registration

• Webflux or Servlet application:
  servlet

Description

The clustering is working fine and applications are being clustered on all servers. The problem is that if a new version of a given application is deployed, then the old instance is never removed from Spring Admin.

yaml

hazelcast:
cluster-name: spring-admin
metrics:
enabled: false
management-center:
enabled: false
retention-seconds: 30
jmx:
enabled: false
map:
spring-boot-admin-event-store:
backup-count: 1
in-memory-format: OBJECT
merge-policy:
batch-size: 100
class-name: PutIfAbsentMergePolicy
time-to-live-seconds: 30
eviction:
eviction-policy: NONE
spring-boot-admin-sent-notifications:
backup-count: 1
in-memory-format: OBJECT
merge-policy:
batch-size: 100
class-name: PutIfAbsentMergePolicy
time-to-live-seconds: 30
eviction:
eviction-policy: LRU
network:
reuse-address: true
outbound-ports:
- 33000-35000
port:
auto-increment: false
port: 5701
join:
auto-detection:
enabled: false
multicast:
enabled: false
tcp-ip:
enabled: false
aws:
enabled: true
hz-port: 5701
service-name: dev-admin-server
interfaces:
enabled: true
interfaces:
- 10.171.0-240.*

Review the supplied Hazelcast clustering and map configuration alongside Spring Boot Admin's self-registration behavior; no source file, test, or entry point is named. Reproduce the deployment of a new application version in the clustered setup and determine why the old instance remains. Done means obsolete instances are removed from Spring Admin after an application replacement.