This is part of our commercial offering.
If you are here, we assume that you have access to a Polyaxon EE Control Plane or Polyaxon Cloud.
- Agents interact with Polyaxon Control Plane to check the queues that they manage for operations to start, stop, or update.
- Agents can deploy different artifacts store and connections to ensure isolation, it's also possible to deploy similar configurations on different clusters to enable access to new resources, e.g. on-prem GPU and GKE for TPUs.
- Agents can be used to deploy jobs / distributed jobs workload, but they can also deploy and expose services.
It's important to note that Polyaxon Agent is an important piece in Polyaxon's architecture to enable complete isolation of workload and artifacts and a hybrid execution and management of resources.
Agents are deployed and use a namespace to run operations independently of other applications running on your cluster.
kubectl create namespace polyaxon # namespace "polyaxon" created
If you would like to use a different value, you must keep in mind to update the
namespace value in your config.
This section will help you create a configuration file to deploy an agent. Polyaxon Agent ships with default values, however and depending on your use case you may need to override some of these values. To do so, you need to create a configuration file and we recommend to save it somewhere safe so that you can reuse it in the future.
Create a config file
and set up all information you want to override in the default config.
If you are using Polyaxon CLI to deploy Polyaxon, we suggest that you define the version and deployment type:
deploymentChart: agent deploymentVersion: 1.1.0
In Polyaxon UI you can get an starter deployment config:
To manage agents please check the management section
First of all, you need to add the Polyaxon helm repository to your Helm, so you can install Polyaxon from it. This makes it easy to refer to the Polyaxon chart without having to use a long URL each time.
helm repo add polyaxon https://charts.polyaxon.com helm repo update
You can validate that your deployment
config.yaml file is compatible with the version you are trying to deploy:
polyaxon admin deploy -f config.yaml --check
To perform a dry run:
polyaxon admin deploy -f config.yaml --dry-run
Now you can install Polyaxon Agent with your
Note: it's important to know that there's an initial delay before you can access Polyaxon API, and before some pods will turn green, which is set to 2 minutes.
You can use Polyaxon CLI to manage the deployment
polyaxon admin deploy -f config.yaml
Or you can use Helm to do the same:
in Helm 2
helm install polyaxon/agent \ --name=<RELEASE_NAME> \ --namespace=<NAMESPACE> \ -f config.yaml
in Helm 3
helm install <RELEASE_NAME> polyaxon/agent \ --namespace=<NAMESPACE> \ -f config.yaml
name is an identifier used by helm to refer to this deployment.
You need it when you are changing the configuration of this install or deleting it.
We recommend using
RELEASE_NAME = polyaxon or
RELEASE_NAME = plx.
--namespace should be the same value of the namespace you created in the first step,
we again recommend using
polyaxon to make it always easy to remember.
Tip: We recommend using
polyaxonfor both the
--namespaceto avoid too much confusion. The same command with
polyaxonas a value:
helm install polyaxon/agent \ --name=polyaxon \ --namespace=polyaxon \ -f config.yaml
Note: "Release name already exists error" If you get a release named
<RELEASE_NAME>already exists error, then you should delete the release by running
helm delete --purge <RELEASE_NAME>.
You can see the pods being created by entering in a different terminal:
kubectl --namespace=<NAMESPACE> get pod
When helm is done deploying Polyaxon, it will output some instructions
these note will be different depending on your configuration (the service type used and/or ingress);
These notes are important for setting the CLI, and getting access to the dashboard.
Next step you need the Polyaxon CLI installed, and you need to configure the host and the ports based on these notes.
To upgrade Polyaxon to a newer version, you can simply run the following command using Polyaxon CLI:
polyaxon admin upgrade -f config.yaml
Or using Helm
helm upgrade polyaxon polyaxon/agent -f config.yaml
When you are done with an agent, you can turn off the deployment, and depending on your persistence configuration you can keep all your data saved for future deployments.
You can also decide to completely turn off Polyaxon and remove the namespace and computational resources.
polyaxon admin teardown -f config.yaml
helm del --purge polyaxon
You can also delete the agent from the dashboard.
Delete the namespace the agent was installed in. This deletes any disks that may have been created to store user’s logs, and any IP addresses that may have been provisioned.
kubectl delete namespace <your-namespace>
If you used the default values, the command should be,
kubectl delete namespace polyaxon