Requirements and supported entities

Before enabling Infrastructure 360 for AWS, ensure you have:

• Auto-discovery enabled for the AWS account.
• The cloudformation:StartResourceScan permission, which the CloudFormation stack grants during setup.
• Resources in a region that Infrastructure 360 scans (see scan frequency below).

What auto-discovery covers on AWS

Auto-discovery scans your connected AWS accounts and creates a New Relic entity for every supported resource, then works out how those resources connect to each other and records those connections as relationships. Auto-discovery doesn't require an agent. Resources appear whether or not they're instrumented.

Auto-discovery creates entities and relationships, not telemetry. A discovered resource gets an entity, tags, its AWS configuration, and a place on your maps, but auto-discovery doesn't send metrics or events for it. To get charts and alerts, connect AWS metrics or install an agent as well.

If you already monitor a resource another way (for example, the infrastructure agent on an EC2 instance, or CloudWatch Metric Streams on an RDS instance), auto-discovery adds to that existing entity rather than creating a duplicate.

Auto-discovery supports 66 AWS resource types, mapping to 65 entity types.

Hosts

Auto-discovery maps the following host entity type:

This is the same entity type the infrastructure agent uses, so an agent-monitored instance stays a single entity.

Serverless

Auto-discovery maps the following serverless entity type:

Containers

Auto-discovery maps the following container entity type:

Compute and containers

Auto-discovery maps the following compute and container entity types:

Auto-discovery covers both ECS launch types (tasks on EC2 and on Fargate) and discovers tasks only while they're running.

Networking

Auto-discovery maps the following networking entity types:

Application, Network, and Gateway Load Balancers share one AWS resource type. Auto-discovery resolves the entity type from the load balancer's ARN. A Gateway Load Balancer becomes an ELB application load balancer entity, since it has no dedicated type of its own.

Databases and caching

Auto-discovery maps the following database and caching entity types:

Which entity type an RDS resource becomes depends on the database engine. DocumentDB engines produce DocumentDB entities. ElastiCache splits into seven entity types based on engine and whether the cache is serverless. A cache cluster becomes a node entity, not a cluster entity, with one node entity per cache node. Valkey replication groups use the ElastiCache Redis Replication Group entity type, so there's no separate Valkey replication group entity. Shards aren't entities. For a cluster-mode replication group, auto-discovery records the shard as a tag on each node.

Messaging and streaming

Auto-discovery maps the following messaging and streaming entity types:

Auto-discovery discovers managed Kafka brokers and topics as entities, but it doesn't currently draw relationships between an MSK cluster and its brokers or topics. The cluster's own relationships (VPC, Subnet, Security Group, KMS key) are unaffected.

Storage and search

Auto-discovery maps the following storage and search entity types:

OpenSearch and Elasticsearch domains share one entity type, so you can't tell them apart by type alone.

Delivery, APIs, and security

Auto-discovery maps the following delivery, API, and security entity types:

Auto-discovery supports WAFv2 only, not WAF Classic. A web ACL is either global scope (protecting CloudFront) or regional scope (protecting API Gateway, ALB, and similar). Auto-discovery registers CloudFront distributions and global-scope WAFv2 web ACLs without an aws.region tag, since they're global resources. Filtering or grouping by region won't return them.

Other

Auto-discovery maps the following additional entity types:

Relationships

As well as creating entities, auto-discovery works out how your resources connect and records those connections as relationships. Relationships are what your maps draw, and what lets you trace a slow request from a load balancer down to the database behind it.

Every relationship consists of a source, a relationship type, and a target. For example, VPC CONTAINS Subnet or S3 bucket TRIGGERS Lambda function. The order matters: it tells you which way the connection goes.

There are eight relationship types:

Auto-discovery records most relationships in one direction only, since many AWS connections have no natural direction (a VPC and a subnet are simply related). Maps still draw the connection correctly from either end, regardless of how auto-discovery recorded it.

Auto-discovery reads most relationships directly from a resource's configuration, but in some cases infers them. For example, it connects a load balancer target group to the right host when the target group routes to IP addresses rather than named instances, by matching target IPs against your network routing.

Scan frequency

Auto-discovery rescans a region based on your account's metrics polling interval, with a default frequency of 24 hours. You can configure rescanning to run every 8, 12, or 16 hours instead. Minimum interval length is 8 hours. If you set an interval below 8 hours, it automatically defaults to 8 hours.

Next steps

Once you've confirmed these prerequisites, see Set up Infrastructure 360 for AWS to enable the integration.