Polyaxon supports scaling of it's services (api and workers) in horizontal way using replication, and the user can increase the workers' concurrency for higher throughput.
To replicate the platform or one of the services (api or workers), you just need to modify the
replicas field of that service you want to scale horizontally.
For example to replicate
api, your config yaml file should include:
api: replicas: 3
This will create 3 pods for api to handle the traffic coming from the CLI, Dashboard, or REST API.
The easiest way to replicate the platform is to increase the
replicas of all Polyaxon's services:
api: replicas: 3 scheduler: replicas: 3 hpsearch: replicas: 3 eventsHandlers: replicas: 3 eventMonitors: replicas: 3
Replication might be easier to scale Polyaxon, but it comes at a memory cost, as it's not always efficient, Polyaxon provides a way to scale it's services' concurrency as well, the rule of thumb is to set the concurrency of the worker you wish to scale to the number of cores available. This will allow to reduce the memory footprint on your cluster and allow the worker to consume more events/tasks.
For example you may want to increase the concurrency of the scheduler:
scheduler: replicas: 2 concurrency: 10
This will create 2 replicas for the scheduler, with 10 concurrent processes each.
By default Polyaxon does not set limits on the resources for the core components it deploys, in order to enable the resources limits, your config yaml file should include:
This will force the Polyaxon to set the resources limits on all services if they include the limits subsections.