- Fabio - @ticofab
- Adam Sandor - @adamsand0r
- Middle meeting - Reactive Distriubted Systems
- Akka : Implementation
- Kubernetetes: Distrubtions
- Can kubernetes make an reactive application?
- Fault toleerence - Component is being hit but keeps going - possibly wiht reduced functionality
- Resilience - Upon failure a component is able to reset into functional state
- Reactive principles -> Reactive design patterns - Reactive building blocks [jvm/application level]
- Actors gossip to each other about what/who they are
- Supervisor can handle situations
- Container orchestration engine at heart
- Has rules for what services have to be up
- Reactive?
- Service abstraction + load balencing
- Simple processing requests
- Pipelines?
- Statemeful microservice architecture
- Sends state to a persistent store
Are actors microservices?
Yes
Actor in akka cluster & microservice on kubernettes
They all have location transparency, resilency, scalability
But they're not independently deployable.
- All code is in one cluster
- Cluster per pod
- Put them all in one akka cluster (communication between pods)
- Mulitple webservice gateways
- Seperate services via clusters
- K8s creates nodes and help form cluster?
- Membership
- Akka Bootstrap (cluster registration)
- Pod restarts based on healthchecks
- System level -Can catch non expected cases
- How does kubernettes handle processing requests that require state?
- Processing/clustering management vs akka
- Auto scaling akka clustering based on demand?
- How do you deploy new code bases to clusters (Managmenet of new code deployments/reprocessing) 4.5 Api management /akka cluster membership
- Scaling down