Requirements

Immuta comprises three core services: Secure, Discover, and Detect. These services rely on PostgreSQL and Elasticsearch to store their states, a caching layer, and Temporal for job execution. The illustration below shows the relationships among these services.

The Immuta Enterprise Helm chart (IEHC) does not include the deployment of PostgreSQL or Elasticsearch, so you must deploy them separately.

Although Immuta recommends using Elasticsearch because it supports several new Immuta features and services, you can deploy Immuta without Elasticsearch. The table below outlines the Immuta features supported with and without Elasticsearch and the dependencies you must deploy and manage yourself.

	Immuta with Elasticsearch	Immuta without Elasticsearch
Dependencies	Externalized PostgreSQL Elasticsearch / OpenShift	Externalized PostgreSQL
Immuta Detect	✅	❌
Audit of Immuta and data platform events	✅	❌
Legacy audit	✅ (Disabled by default, but can be enabled)	✅ (Until October 2024)
Immuta Monitors	✅	❌
Sensitive data discovery	✅	✅

For more information about legacy features and services no longer enabled in the recommended deployment of Immuta, see the Legacy features and services section.

Version requirements

Kubernetes versions

• Kubernetes 1.29 to 1.32

Metadata database (PostgreSQL)

PostgreSQL incompatibilities

Immuta is not compatible with PostgreSQL abstraction layers, such as Amazon Aurora.

• PostgreSQL 15.0 or newer

• The pgcrypto, btree_gin extensions must be enabled

Elasticsearch

• Elasticsearch v7 API or newer

• OpenSearch compatible with Elasticsearch v7 API or newer

OpenSearch user

The user provided during the install must have the following permissions:

• cluster:monitor/health

• indices:data/write/bulk*

• indices:data/write/bulk

• indices:data/read/search

• indices:admin/exists

• indices:admin/create

• indices:admin/delete

• indices:admin/settings/update

• indices:admin/get

• indices:data/write/delete/byquery

• indices:data/write/index

• indices:admin/mapping/put

• indices:data/write/bulk

• indices:data/write/bulk*

Follow OpenSearch documentation to create the user and add permissions, or see the Setting up OpenSearch permissions knowledge base article.

Cache (Redis/Memcached)

Built-in cache

The IEHC manages its own Memcached deployment inside the cluster. The key-value cache can optionally be externalized post installation.

• Redis 7.0 or newer

• Memcached 1.6 or newer

Temporal

Built-in Temporal server

The IEHC deploys a Temporal server and its requisite components. However, you may choose to use your own Temporal instance.

• Temporal 1.24.2 or newer

Infrastructure recommendations

Kubernetes distribution	Ingress	External metadata database	External Elasticsearch
Amazon Elastic Kubernetes Service (EKS)	AWS Load Balancer Controller	Amazon RDS for PostgreSQL	Amazon OpenSearch
Azure Kubernetes Service (AKS)	Azure Application Gateway Ingress Controller	Azure Database for PostgreSQL	Elastic Cloud on Azure
Google Kubernetes Engine (GKE)	GKE Ingress Controller	Google Cloud SQL for PostgreSQL	Elastic Cloud on Google Cloud
Red Hat OpenShift	OpenShift Ingress Operator

Legacy features and services

Some legacy services and features are no longer enabled in the recommended configuration of the IEHC. The table below lists these features and provides links to documentation that outlines how to enable them in Immuta.

Feature	Immuta Enterprise Helm chart configuration
Legacy audit	Set each of the following global.featureFlags in your immuta-values.yaml file to false: AuditService detect auditLegacyViewHide
Legacy conditional tags	Set the following global.featureFlags in your immuta-values.yaml file to false: DiscoverDeprecateLegacyTags
Legacy sensitive data discovery	Enable the query engine and fingerprint services
Data platforms Amazon Redshift Azure Synapse Analytics Google BigQuery Legacy databases	Enable the query engine
Policies Masking with format preserving masking (unless using the Snowflake integration) Masking with k-anonymization Masking using randomized response (unless using the Snowflake integration)	Enable the query engine and fingerprint services