Deployment: Kubernetes-Based

This page gives you an overview of the Pulse Kubernetes-Based deployment.

You can either give a dedicated namespace in your existing Kubernetes for Pulse, or you can provision a completely new Kubernetes cluster.

Deploy Agents, Hooks, and Connectors

Agents

Deploy Pulse Agents on Hadoop cluster nodes to collect logs, traces, metrics, and other telemetry data.

You can install Agents using one of the following methods:

• Accelo Binary – Use this method if SSH is enabled on the cluster nodes.
• MPacks or Parcels – Use MPacks for installation via Ambari, or Parcels via Cloudera Manager.
• Hystaller – Use this method for offline Agent installation.

Hooks

Place Hook JARs on Hadoop nodes to collect Hadoop service query events.

Connectors

Deploy Connectors as Add-on services in the Pulse Server to fetch, process, and analyze Hadoop service data.

For details about Agents, Hooks, and Connectors, see Pulse Agents, Pulse Hooks, and Connectors.

Install Pulse Manager Server

The Pulse Manager Server is an API-based service used to manage the deployment of Pulse Operator and Pulse Database services.

You can install the Pulse Manager Server along with the Pulse CRD, Pulse Operator, and Pulse Manager Agent using the installation package provided by Acceldata.

• Pulse Manager Agent deploys Pulse databases on virtual machines (VMs).
• Pulse Operator deploys Pulse services within Kubernetes cluster nodes.

For details about installing Pulse Manager Server, see Install Pulse Manager Server.

Deploy Pulse Services in Kubernetes

This is the workflow for deploying Pulse services in a Kubernetes cluster:

• The Pulse Manager Server generates a Custom Resource from the user-provided cluster configuration.
• This Custom Resource is applied to the Pulse Operator.
• The Pulse Operator creates the necessary Kubernetes resources such as Deployments, StatefulSets, and Services.
• These resources deploy the required Pulse services in the cluster.

For details about deploying Pulse Core and Add-on services, see Deploy Pulse Core Services and Deploy Pulse Add-on Services.

For details about Pulse Core and Add-on services, see Pulse Core Services and Pulse Add-on Services.

Deploy Data Storage Services in VMs

This is the workflow for deploying Pulse databases on VMs or bare-metal servers:

• Deploy Pulse databases on virtual machines (VMs) or bare-metal servers.
• The Pulse Manager Server coordinates the database deployment configuration with the Pulse Manager Agent.
• Install the Pulse Manager Agent on each VM to deploy the databases and report deployment status.
• Use an odd number of VMs (e.g., 1, 3, 5, or 7) to ensure performance and high availability (HA).
• Deploy at least three VMs to ensure fault tolerance and load distribution.

For details about Pulse Databases, see Pulse Databases.

For details about deploying Pulse databases, see Deploy Pulse Database.

Clustered Databases

Pulse uses clustered databases to ensure high availability, scalability, and efficient data management across all VMs.

MongoDB (Sharded Cluster)

• Deployed on all VMs. Stores query and analytics data, Hadoop, and Pulse configuration data
• Components: ad-db-router, ad-db-config, ad-db-shard

ElasticSearch Cluster

• Deploy on all VMs as master-eligible and data nodes. Stores log data.
• Component: ad-elastic

VictoriaMetrics Cluster

• Deploy on all VMs. Stores time-series metrics.
• Components: ad-vminsert, ad-vmselect, and ad-vmstorage

NATS Cluster

• Deploy on all VMs. Handles event messaging.
• Component: ad-events

Supporting Database Services

These lightweight services support visualization, log processing, and storage optimization.

PostgreSQL

• Deploy on one node. Stores Dashplot metadata.
• Components: (ad-pg and ad-pg-ui)

LogStash

• Deploy on all nodes. Forwards logs to ElasticSearch.

Gauntlet

• Deploy on one node. Performs database cleanup based on retention policies.

For detailed

Summary

• Kubernetes handles orchestration, processing, automation, and UI delivery.
• VMs handle database deployments for telemetry and storage.
• Pulse Manager Server and Pulse Operator ensure seamless coordination across environments.
• Clustering ensures high availability and performance for all critical data flows.

For detailed steps on deploying Pulse on Kubernetes, see Kubernetes-Based Installation.