Scaling Microservices

Scaling microservices

You can instruct the tool to run multiple instances of any of the following microservices.

• autoignore

• autostarter

• configuration

• database

• detection

• fileobserver

• mitigation

• notifier

• prefixtree

• riperistap

• bgpstreamlivetap

• bgpstreamkafkatap

• bgpstreamhisttap

• exabgptap

However, we recommend not scaling tap microservices, and in case you have issues with large BGP update loads, to scale only the following microservices:

• detection (for load-balancing detection processes)

• prefixtree (for load-balancing prefix tree lookups; attention, this requires memory!)

• database (for load-balancing database access)

To apply scaling (can be upwards or downwards):

• Before you initiate artemis:

  docker-compose up --scale <service_1>=number_of_service_1_replicas> --scale <service_2>=number_of_service_2_replicas> ... -d
  

• While ARTEMIS is running (live):

  docker-compose scale <service_1>=number_of_service_1_replicas>
  docker-compose scale <service_2>=number_of_service_2_replicas>
  ...