J4K-2019: Paul Parkinson: Helidon and Kubernetes
Raw Notes from Paul Parkinson’s session.
Workflow of Getting into Kubernetes
-
Service Broker, Kubernetes Secret, Microservice
-
CloudEvents
standard -
ServiceMesh, Istio, Envoy Data Plane
-
Side note: Tracing has become easier.
Kiali
. Uses opentracing, Jaeger, Zipkin.- Gave some background on Jaeger.
-
WebLogic Kubernetes Operator
-
Helidon SE
-
Microframework
-
Functional
-
Reactive
-
Tracing
-
No annotations
-
Messaging client
-
-
Helidon MP
-
Full Microprofile standard
-
Familiar to Java EE Devs
-
Tracing
-
JTA/JPA
-
Oracle DB feature integration
-
Long Running Activities
-
Long Running Activities
-
Annotations
@LRA
. These are the newer version of JTA -
Ties in to DDD Bounded Contexts. When decomposing the monolith and the DB in particular, this is where migration efforts fail. Close attention needs to be paid.
Microservices Data Management Patterns
-
CQRS
-
Access thru public APIs onlt
-
Database per service
Microservices Communication
-
No locking
-
Event based integration
-
Pub/sub
-
Event Sourcing
- Gave basic definition, illustrated how it works.
-
CQRS
- Same story, gave definition and illustrated how it works.
-
Atomicity of Persisting Events and Database States
- Oracle Advanced Queue: A messing system in the database. 16 year old technology, but fits really well with Microservices. Guaranteed once delivery. Idempotency for free.
-
SAGA Pattern
-
A saga is a sequence of local transactions. Two flavors. Choreography and Orchestration. The latter is you have an orchestrator that controls the actions of the participants.
-
Compensating transactions. Showed some code slides of complexity with and without AQ.
-
-