AgentCore Observability Will Take Your Traces From Anywhere, If You Build the Pipe

Claim ledger

Ranked by verification strength, evidence, and original report placement.

1. 01
   [1]

   AWS published a post on the AWS Machine Learning Blog titled "Monitor on-premises and multi-cloud AI agents with AgentCore Observability", showing how to set up observability for agents running outside AWS, configure AWS Distro for OpenTelemetry auto-instrumentation in non-AWS environments, route telemetry to the AgentCore Observability dashboard, and validate the setup end to end.

   Reported·Source: AWS Machine Learning BlogView cited source
2. 02
   [2]

   The post states that agents built with frameworks like Strands Agents, LangGraph and CrewAI need performance observability whether they run on Amazon EKS, Amazon ECS, AWS Lambda, on-premises, or another cloud provider such as Google Cloud Platform or Microsoft Azure.

   Reported·Source: AWS Machine Learning BlogView cited source
3. 03
   [3]

   Amazon Bedrock AgentCore Observability provides native tracing, monitoring and analytics, but natively supports only agents deployed on the AgentCore runtime in the AWS Cloud; if agents run anywhere else, additional configuration is required to send telemetry to the dashboard.

   Reported·Source: AWS Machine Learning BlogView cited source
4. 04
   [4]

   The described solution runs AWS Distro for OpenTelemetry (ADOT) in-process with the agent application; ADOT auto-instruments the agent framework, captures generative AI semantic convention spans, and exports telemetry directly to the Amazon CloudWatch OpenTelemetry Protocol (OTLP) endpoint using SigV4 authentication with AWS IAM credentials.

   Reported·Source: AWS Machine Learning BlogView cited source
5. 05
   [5]

   Sending telemetry from an AI agent to AgentCore Observability requires three core components: ADOT auto-instrumentation, IAM credentials used as access keys to authenticate with CloudWatch and forward traces, metrics and logs, and environment variables containing specific OpenTelemetry settings for routing and authentication.

   Reported·Source: AWS Machine Learning BlogView cited source
6. 06
   [6]

   The prerequisites specify IAM user credentials (access key ID and secret access key) with permissions for the listed actions.

   Reported·Source: AWS Machine Learning BlogView cited source