AI Engine (DCIE)

How long it takes to generate the first recommendation?
If a pod starts after Federator.ai deployment, Federatora.ai generates the first recommendation within 10 minutes after the pod starts. If a pod is started before Federator.ai deployment and has enough accumulated historical data, Federator.ai generates recommendations without waiting for data collection. Federator.ai gives out recommendations every 5 minutes.

Basic

What is Federator.ai for OpenShift exactly?
Federator.ai for OpenShift (Federator.aifor short) is a new generation AIOpssolution designed explicitly for OpenShift environment. Its main goal is to optimize container resource configurations and simplify operations for users running OpenShift. After deploying Federator.aiin your OpenShift cluster,
it forecasts resource demands of containerized applications based on historical data, recommends just-fitted pod sizes and replicas based on the prediction, re-configures pod sizes and replicas with the right values at the right time.
What are the main features?
Its main features include:
* Performance metrics prediction
– Per-container workload (CPU and memory) prediction
– Node resource usage (CPU and memory) prediction
* Policy-driven resource recommendation, including two policies: Stable and Compact
– Per-container requests and limits values (CPU and memory) recommendation, including when(a time window) to re-configure at what requestsand limitsvalues
– Pod placement recommendation (which node to place a pod)
* Pod re-sizing/placement recommendation execution
* Dashboard
What types of pod scaling are supported by Federator.ai?
It supports prediction-based vertical pod scaling (requests and limits recommendation of a container) in the latest release. In the next release (early June), it will support prediction-based horizontal pod scaling (replicas of a pod).
What are the target environments?
You can install Federator.ai in any OpenShift environment, no matter it is deployed on-premises or public clouds.
Is there a free trial?
Yes, we provide a free trial for both on-premises and public cloud users.
What is your business model?
We offer SaaS (software as a service) model. Users need a subscription and acquire a license. Currently, we offer free-trial for all users without requiring a license keycode. We will add license keycode in the future release.

Federator.ai Operator

What is Federator.ai Operator?
Federator.ai Operator is a Kubernetes native application created on Operator Framework. After Federator.ai is installed with Federator.ai Operator, Federator.ai Operator is responsible for managing the upgrade and lifecycle of Federator.ai.
Where to find Federator.ai Operator?
You can find it on OperatorHub.io. OperatorHub.iois a public registry for operators and operated by Red Hat in collaboration with AWS, Microsoft Azure, and Google Cloud. Another method is from Red Hat Container Catalog. Both places have the same container images.
Why created Federator.ai Operator?
With Federator.ai Operator listed OperatorHub.io, users can expect that Federator.ai has several features that are important for production environments. Currently, our Federator.ai has Level-3 capability, including (1) easy installation, (2) seamless upgrade (rolling upgrade of containers), and (3) lifecycle management (liveness and readiness check, storage provisioning through PVC (persistent volume claims).
What happens after Federator.ai Operator is deployed in my OpenShift environment?
Federator.ai Operator pulls the following containers from Quay.io and deploys them in your OpenShift environment. Quay.io is a secured Docker image hosting service operated by Red Hat.
Do I need to install a new Prometheus for running Federator.ai?
Federator.ai can utilize current Prometheus installation. If there is already a Prometheus in your K8s Cluster, you can provide its connection info to Federator.ai. Federator.ai will collect pod and node metrics from the existing Prometheus.

Federator.ai Operator doesn’t provide Prometheus installation, and you have to install it yourself.
Can I use Federator.ai Operator in vanilla Kubernetes?
Yes, the installation process is very similar to OpenShift installation. The only difference is that the command line is the only method for vanilla K8s, it doesn’t have equivalent OpenShift 4.0 console.

Installation

How to install?
There are two approaches:
(1) Install on OKD by OLM (Operator-lifecycle-management)
  1. Install Federator.ai operator via OLM
  2. Create Federator.ai service via Federator.ai operator AlamedaServiceCRD
(2) Install on vanilla K8s by OLM
  1. Install OLM framework if it is not installed yet
  2. Install Federator.ai operator via OLM
  3. Create Federator.ai service via Federator.ai operator AlamedaServiceCRD
For more information about OLM, please refer to its Github repository.
How to install with OLM by the command line?
There are two commands to install Federator.aiby using Federator.ai Operator
(1) Step 1 to install prerequisites
$ kubectlcreate -f https://raw.githubusercontent.com/operator-framework/operator-lifecycle-manager/master/deploy/upstream/quickstart/olm.yaml
(2) Step 2 to the operator
$ kubectlcreate -f https://operatorhub.io/install/alpha/federatorai.v0.0.1.yaml
Share