Cluster IQ

Cluster IQ is a tool for making stock of the Openshift Clusters and its resources running on the most common cloud providers and collects relevant information about the compute resources, access routes and billing.

Metrics and monitoring is not part of the scope of this project, the main purpose is to maintain and updated inventory of the clusters and offer a easier way to identify, manage, and estimate costs.

ClusterIQ has a Web UI called ClusterIQ Console. Follow this link for installation instructions.

Supported cloud providers

The scope of the project is to cover make stock on the most common public cloud providers, but as the component dedicated to scrape data is decoupled, more providers could be included in the future.

The following table shows the compatibility matrix and which features are available for every cloud provider:

Cloud Provider	Compute Resources	Billing	Actions	Scheduled Actions
AWS	Yes	Yes	Yes	Yes
Azure	No	No	No	No
GCP	No	No	No	No

Architecture

The following graph shows the architecture of this project:

Documentation

The following documentation is available:

• Events Documentation - Event flows and sequence diagrams
• Development Setup - Local development guide

Installation

This section explains how to deploy ClusterIQ and ClusterIQ Console.

Prerequisites:

Cloud provider RBAC configuration

Before configuring credentials for ClusterIQ, it is recommended to access the user and permission management service and create a dedicated user exclusively for ClusterIQ. This user should have the minimum necessary permissions to function properly. This approach enhances the security of your public cloud provider accounts by enforcing the principle of least privilege.

Each Cloud Provider has a different way for configuring users and permissions. Before continuing, check and follow the steps for each Cloud Provider you want to configure:

• Amazon Web Services (AWS)
• Microsoft Azure Not available.
• Google cloud Platform: Not available.

Accounts Configuration

1. Create a folder called secrets for saving the cloud credentials. This folder is ignored on this repo to keep your credentials safe.

   mkdir secrets
   export CLUSTER_IQ_CREDENTIALS_FILE="./secrets/credentials"
   

   ⚠️ Please take care and don't include them on the repo.

2. Create your credentials file with the AWS credentials of the accounts you want to scrape. The file must follow the following format:

   echo "
   [ACCOUNT_NAME]
   provider = {aws/gcp/azure}
   user = XXXXXXX
   key = YYYYYYY
   billing_enabled = {true/false}
   " >> $CLUSTER_IQ_CREDENTIALS_FILE
   

   ⚠️ The values for provider are: aws, gcp and azure, but the scraping is only supported for aws by the moment. The credentials file should be placed on the path secrets/* to work with docker/podman-compose.

   ❗ This file structure was design to be generic, but it works differently depending on the cloud provider. For AWS, user refers to the ACCESS_KEY, and key refers to SECRET_ACCESS_KEY.

   ❗ Some Cloud Providers has extra costs when querying the Billing APIs (like AWS Cost Explorer). Be careful when enable this module. Check your account before enabling it.

Openshift Deployment

Since version 0.3, ClusterIQ includes its own Helm Chart placed on ./deployments/helm/cluster-iq. For more information about the supported parameters, check the Configuration Section.

1. Prepare your cluster and CLI

   oc login ...
   
   export NAMESPACE="cluster-iq"
   oc new-project $NAMESPACE

2. Create a secret containing this information is needed. To create the secret, use the following command:

   oc create secret generic credentials -n $NAMESPACE \
     --from-file=credentials=$CLUSTER_IQ_CREDENTIALS_FILE

3. Configure your cluster-iq deployment by modifying the ./deployments/helm/cluster-iq/values.yaml file.

4. Deploy the Helm Chart

   helm upgrade cluster-iq ./deployments/helm/cluster-iq/ \
     --install \
     --namespace $NAMESPACE \
     -f ./deployments/helm/cluster-iq/values.yaml

5. Monitor every resource was created correctly:

   oc get pods -w -n $NAMESPACE
   helm list -n $NAMESPACE

6. Once every pod is up and running, trigger the scanner manually for initializing the inventory

   oc create job --from=cronjob/scanner scanner-init -n $NAMESPACE

Uninstalling

To uninstall ClusterIQ Helm chart, use the following commands

helm uninstall cluster-iq -n $NAMESPACE
helm list -n $NAMESPACE

Local Deployment (for development)

For deploying ClusterIQ in local for development purposes, check the following document

DB Backup

For backing up or restoring the ClusterIQ database, check the following document This document also describes how to manage data migration when a new release of ClusterIQ changes DB data structure.

Configuration

Available configuration via Env Vars:

Key	Value	Description
CIQ_AGENT_INSTANT_SERVICE_LISTEN_URL	string (Default: "0.0.0.0:50051")	ClusterIQ Agent gRPC listen URL
CIQ_AGENT_POLLING_SECONDS_INTERVAL	integer (Default: 30)	ClusterIQ Agent polling time (seconds)
CIQ_AGENT_URL	string (Default: "agent:50051")	ClusterIQ Agent listen URL
CIQ_API_LISTEN_URL	string (Default: "0.0.0.0:8080")	ClusterIQ API listen URL
CIQ_API_URL	string (Default: "")	ClusterIQ API public endpoint
CIQ_AGENT_LISTEN_URL	string (Default: "0.0.0.0:50051")	ClusterIQ Agent listen URL
CIQ_DB_URL	string (Default: "postgresql://pgsql:5432/clusteriq")	ClusterIQ DB URL
CIQ_CREDS_FILE	string (Default: "")	Cloud providers accounts credentials file
CIQ_LOG_LEVEL	string (Default: "INFO")	ClusterIQ Logs verbosity mode
CIQ_SKIP_NO_OPENSHIFT_INSTANCES	boolean (Default: true)	Skips scanned instances without cluster

Scanner

The scanner searches each region for instances (servers) that are part of an Openshift cluster. As each provider and each service has different specifications, the Scanner includes a specific module dedicated to each of them. These modules are automatically activated or deactivated depending on the configured accounts and their configuration.

# Building in a container
make build-scanner

# Building in local
make local-build-scanner

API Server

The API server interacts between the UI and the DB.

# Building in a container
make build-api

# Building in local
make local-build-api

Agent (gRPC)

The Agent performs actions over the selected cloud resources. It only accepts incoming requests from the API.

Currently, on release v0.4, the agent only supports Power On/Off clusters on AWS.

# Building in a container
make build-agent

# Building in local
make local-build-agent