Migration and Transfer
AWS Application Discovery Service
On the radar· Application Discovery Service no longer accepts new customersCollects usage and configuration data about on-premises servers and databases to plan a migration to AWS, for example to group servers into applications based on their network connections. It has not accepted new customers since November 7, 2025; AWS points to AWS Transform as the alternative.
Exam phrases that point here
«collect information about on-premises servers to plan the migration to AWS»
Application Discovery Service inventories on-premises servers and databases (configuration, CPU, memory, and disk usage) to decide what to migrate and how to size it on AWS. Note: since November 7, 2025 it no longer accepts new customers, and AWS suggests AWS Transform for new projects.
When not: Discovering is not migrating: if the scenario asks to move the servers to EC2, it points to AWS Transform MGN (formerly AWS Application Migration Service), and if it asks to move the data of a database, to AWS DMS. In a current scenario, the tool AWS recommends for discovery is AWS Transform.
«identify the network dependencies between servers before migrating them»
The Discovery Agent records each server's inbound and outbound network connections; with them you can see which servers talk to each other and group them into applications to migrate together.
When not: The VMware Agentless Collector does not see the processes running inside each VM. If you only need resource usage and nothing must be installed on the servers, the agentless collector is enough; if the goal is to monitor network traffic already in AWS, think of VPC Flow Logs, not this service.
«discover the VMware vCenter virtual machines without installing an agent on each one»
The Agentless Collector is deployed as an appliance (OVA) in vCenter and from there it discovers the hosts and VMs, with their configuration and their average and peak usage.
When not: If there are physical servers or you need to know which processes run inside each machine, the agentless collector is not enough and the Discovery Agent is installed on each server.
«estimate the size and cost of EC2 instances based on the actual usage of on-premises servers»
The performance and usage data the service collects can be exported for a cost model or used to generate EC2 instance type recommendations.
When not: If the servers are already in AWS and the goal is to right-size the instances in use, the clue points to AWS Compute Optimizer.
«query the discovery data of all servers with SQL in one place»
With data exploration turned on, the data from the Discovery Agents is periodically stored in an S3 bucket and queried with Amazon Athena, which comes with predefined queries for performance, processes, and network dependencies.
When not: This continuous export only works with data from the Discovery Agents, not with data from the Agentless Collector. If the source data does not come from discovery, the usual combination is S3 + Athena without this service.