Monitor and Debug Agents with Mastra Observability

​When your agent fails in production, “it worked on my machine” is not a debugging strategy. You need visibility into what happened across every model call, tool invocation, workflow step, and deploy.

​In this workshop, you’ll learn how Mastra Observability gives you logs, traces, and metrics for your agents in one place. You’ll see how to search across projects and deploys, inspect traces, understand failures, and spot cost or latency issues before they become user-facing problems.

​You’ll learn how to:

• ​Trace multi-step agent runs from input to final output

• ​Use logs to isolate failures in tools, workflows, and model calls

• ​Read metrics for cost, latency, reliability, and usage patterns

• ​Understand how Mastra Observability uses OTel-shaped signals and ClickHouse under the hood

• ​Set it up with MastraPlatformExporter() or enable it during npm create mastra@latest

​By the end, you’ll know how to add observability to your Mastra projects and where to look when your agents behave unexpectedly.

​Recording and code examples will be available to everyone who registers.

​Hosted by

• ​Alex Booker, Mastra

  • ​https://x.com/bookercodes

  • ​https://mastra.ai

• ​Joel Smith, Mastra

  • ​https://x.com/jsumnersmith

​Recording and code examples will be available to everyone who registers.