Operations | Monitoring | ITSM | DevOps | Cloud

What is a Session Replay? Record video-like user-sessions to debug your app. Sessions Replays show you what happened and when, along with providing you with all of the relevant logs and traces to debug a user issue without them ever needed to write in.

Lets be real, who really knows whats happening inside MCP Server? Well, now you can! We've been building our MCP Server for the past few months, and we've taken all the tools that we've used to debug it and turned them into a set of observability tools to help you monitor your MCP servers end to end.

We just launched MCP server monitoring in beta. You can instrument most server-side JavaScript SDK based MCP servers with one line of instrumentation code within your MCP SDK implementation using: wrapMcpServerWithSentry(McpServer) See details like protocol usage, client usage, traffic, tool usage, and performance across your MCP implementation so you you can get visibility into all the sharp edges that your MCP server has — who’s using it, how it’s working (or not), and get alerted when things break.

Build an MCP server starting from 0 using Vercel's Next.js template, and ends in generating hot takes in the tone of David Cramer. Scaffold Sentry, add tool calls via Cursor CLI, and setup MCP monitoring with Sentry's new wrapMcpServerWithSentry method.

Replays in Sentry are awesome. With one property in your Sentry config you can start capturing video-like replays of user interactions with your application, but the problem is... you still have to watch them... but not anymore! AI replay summaries take your replays and run the events through an LLM to summarize the events that happened in them. They are broken up into chapters, with the breadcrumb sequences embedded in, so you can quickly get context of whats happening in every replay.

What if you had 5 errors, and instead of coming back to 5 issues in your feed, you got 5 pull requests fixing them? Seer is Sentry's new AI Debugging agent. it's able to stitch together all the context from your logs, stack traces, distributed tracing, codebase, and issues and figure out what broke, where, and how to fix it. Seer automation lets you automate that flow - and end up with a nice PR waiting for you to merge if it looks good. Check it out!

Sentry now has logs. Collect and aggregate logs in your Laravel apps, in both the backend and the frontend. Errors won't always capture the whole story, let's see how logs can help us identify a real customer issue.

Sentry has released AI Agent monitoring, and in this video you can see the fast path to getting started with it using the Vercel AI SDK and Anthropic Claude. AI Agent Monitoring uses tracing to let you see details around how AI interactions are happening inside your application. You can see the back and forth conversation flow, token usage, model usage, durations, and much more. Agent Monitoring is out now, take it for a spin, let us know what you think in Discord!

Monitoring agents and LLM applications is... different. Managing everything from tool calls, to model configurations, token usage, and AI systems do their best to solve problems on their own - so errors aren't always clear. Sentry's agent monitoring focuses on making it easy to dive into your AI applications and understand whats breaking, where, so you can fix it faster.