boundary/website/content/docs/concepts/service-discovery.mdx

---
layout: docs
page_title: Service Discovery
description: |-
  An overview of service discovery in Boundary
---

# Service Discovery

## Overview
Traditionally, connecting to remote hosts and services requires knowledge of 
the endpoint’s connection info (e.g. the IP address and port of the service). 
This creates complexity when managing the onboarding of new resources at scale
or dealing with dynamic services whose connection info frequently changes.

**Service discovery** focuses on automating the process of onboarding new or 
changed infrastructure resources – and their connection info – to Boundary 
as hosts.

# Automating Service Discovery in Boundary

Boundary supports target/service discovery in three primary workflows: 

**[Manual configuration](/boundary/tutorials/oss-administration/oss-manage-targets)**: 
Boundary administrators can manually configure static 
hosts and targets via the administrator UI and CLI. Manual configuration of 
targets with static hosts requires knowledge of the IP address or endpoint used 
to connect to a host.

**[Service discovery via configuration as code with Terraform](/boundary/tutorials/oss-getting-started/oss-getting-started-config)**: 
Boundary is fully programmatically instrumented and the discovery and configuration of new 
infrastructure targets can be automated with 
[Boundary’s Terraform provider](https://registry.terraform.io/providers/hashicorp/boundary/latest/docs). 
This allows for dynamic configuration of a host and target without the need 
for prior knowledge of the target’s connection info. 

**[Runtime service discovery via dynamic host catalogs](/boundary/tutorials/access-management/azure-host-catalogs)**: 
Boundary dynamic host catalogs automate the ingestion of resources from 
infrastructure providers into Boundary. Boundary hosts are automatically 
created, updated and added to host sets in order to reflect the connection 
information maintained in these providers. This removes the need to know 
host connection info or reapply infrastructure as code templates to 
configure new or changed resources.

# Dynamic Host Catalogs
Dynamic host catalogs are an agentless workflow for Boundary to 
securely query infrastructure providers at runtime to discover and configure 
new services. Boundary dynamic host catalogs are written in go-plugin and run 
as separate processes. Boundary administrators can define rules for which 
external resources should be ingested into the catalog by 
[creating dynamic host](/boundary/docs/concepts/domain-model/host-sets)
 sets with an attributes filter. Attributes specify the fields which the plugin 
 should use to lookup which hosts should be members of this host set.

Currently, Boundary supports dynamic host catalog implementations for AWS and 
Azure and we will continue to grow this ecosystem to support additional providers. 

You can get started with dynamic host catalogs [here](/boundary/tutorials/access-management/azure-host-catalogs).