# Sentry Blog Product, Engineering, and Marketing updates from the developers of Sentry. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Debugging our AI search assistant with agent tracing DevFeed: [Debugging our AI search assistant with agent tracing]() Original publisher: [Read original article]() Author: Dominik Buszowiecki; Shaun Kaasten Published: 2026-09-11T09:00:00Z Content type: article Language: en Sources: [Sentry Blog]() Topics: [AI search](), [agent observability](), [debugging](), [tracing](), [Large Language Model]() Tags: [agent](), [ai](), [api](), [debugging](), [errors](), [eval](), [llm](), [search](), [sentry](), [tracing]() ### AI overview Sentry engineers describe how they debugged the Search Query Assistant, which converts natural-language prompts into Sentry Syntax queries. They used evals for performance measurement and AI Conversation tracing to investigate failures, including a bug involving custom numerical attributes that caused queries to return no results. ### Source excerpt See how Sentry engineers used AI Conversations to debug a natural language search assistant and fix a tricky query generation bug. ## Using Sentry Application Metrics to improve a browser video size estimator DevFeed: [Using Sentry Application Metrics to improve a browser video size estimator]() Original publisher: [Read original article]() Author: Kyle Tryon Published: 2026-09-01T09:00:00Z Content type: tutorial Language: en Sources: [Sentry Blog]() Topics: [telemetry](), [Transcodings](), [browser](), [Open Source]() Tags: [app](), [browser](), [encoding](), [media](), [metrics](), [open-source](), [video]() ### AI overview This tutorial explains how the author used Sentry Application Metrics to evaluate and improve a video size estimator in Cliparr, a self-hosted browser-based video clipping tool. Real-world conversion data revealed one estimate was off by 83%. ### Source excerpt The numbers your app lives on don't belong in logs or on sampled spans. Here's how a browser video converter's KPIs made the case for Application Metrics. ## From one switch to a control panel: meet \`dataCollection\` DevFeed: [From one switch to a control panel: meet \`dataCollection\`]() Original publisher: [Read original article]() Author: Sigrid Huemer Published: 2026-08-28T09:00:00Z Content type: release Language: en Sources: [Sentry Blog]() Topics: [SDKs](), [data](), [JavaScript](), [sensitive data](), [Release notes](), [passwords]() Tags: [api-keys](), [data](), [javascript](), [passwords](), [release-notes](), [sdks](), [sensitive-data]() ### AI overview Sentry is replacing the all-or-nothing `sendDefaultPii` setting with `dataCollection`, which provides granular control over automatically collected data such as user information, headers, request bodies, and GenAI data. The change is rolling out across Sentry SDKs, with JavaScript SDK v11 making it the default and collecting more data than v10 by default. ### Source excerpt Sentry SDKs replace the `sendDefaultPii` boolean with `dataCollection`, granular options for user data, headers, bodies, GenAI data, and more. ## How to evaluate session replay software: a developer's guide DevFeed: [How to evaluate session replay software: a developer's guide]() Original publisher: [Read original article]() Author: Lindsay Piper Published: 2026-08-26T09:00:00Z Content type: tutorial Language: en Sources: [Sentry Blog]() Topics: [session replay](), [Playback](), [sensitive data](), [debugging](), [Mobile]() Tags: [debug](), [debugging](), [developer](), [engineering](), [guide](), [libraries](), [monitoring](), [mutation](), [playback](), [privacy](), [sdk](), [sensitive-data](), [session-replay](), [tool](), [ux]() ### AI overview A developer guide to evaluating session replay software by comparing recording methods, privacy architecture, error-monitoring integration, overhead, AI debugging compatibility, and mobile support. It distinguishes UX analytics tools from developer debugging tools and explains the tradeoffs between DOM snapshot recording and pixel capture. ### Source excerpt Compare session replay tools on recording methodology, privacy architecture, integration depth, overhead, AI-readability, and mobile support ## Automated agent triage with Agent Tracing and Claude Routines DevFeed: [Automated agent triage with Agent Tracing and Claude Routines]() Original publisher: [Read original article]() Author: Trevor Elkins Published: 2026-08-13T09:00:00Z Content type: article Language: en Sources: [Sentry Blog]() Topics: [Claude](), [AI Agent](), [tracing](), [Model Context Protocol](), [agent observability]() Tags: [agent](), [ai-agent](), [claude](), [mcp](), [model-context-protocol-server](), [report](), [tool](), [tracing]() ### AI overview Sentry describes a Claude Routine that uses the Sentry MCP to query Agent Tracing data from roughly 800 overnight AI agent conversations. It aggregates error statistics, samples traces and conversations, reviews the Seer codebase, and files new Linear tickets for findings that are not already tracked. ### Source excerpt How Sentry uses a Claude Routine and the Sentry MCP to automatically triage 800 AI agent conversations overnight and file bugs. ## What's new in Sentry Logs: The summer 2026 roundup DevFeed: [What's new in Sentry Logs: The summer 2026 roundup]() Original publisher: [Read original article]() Author: Ben Coe; Allison Rogers Published: 2026-08-11T09:00:00Z Content type: release Language: en Sources: [Sentry Blog]() Topics: [Usability](), [tracing](), [data]() Tags: [2026](), [changelog](), [improvements](), [logs](), [usability]() ### AI overview A summer 2026 roundup of Sentry Logs updates, including log pinning, shareable URL state, large-scale exports with JSONL support, aggregate-aware exports, faster loading, and log viewer usability improvements. ### Source excerpt Everything that shipped for Sentry Logs this summer: log pinning, JSONL exports, terabyte-scale search, and a dozen usability improvements. ## How we built an automated debugging workflow at Sentry DevFeed: [How we built an automated debugging workflow at Sentry]() Original publisher: [Read original article]() Author: Dhrumil Parekh Published: 2026-08-06T09:00:00Z Content type: tutorial Language: en Sources: [Sentry Blog]() Topics: [debugging](), [Automation](), [Artificial Intelligence](), [Claude](), [GitHub](), [Server]() Tags: [agent](), [ai](), [automated](), [automation](), [bugs](), [claude](), [debugging](), [github](), [pull-requests](), [review](), [sentry](), [slack](), [tool](), [workflow](), [workflows]() ### AI overview Sentry describes how it uses Seer, an AI-powered debugging agent, to detect issues, identify root causes, generate fixes, open GitHub pull requests, and route work for human review. The article also explains how organizations can control the level of automation. ### Source excerpt How Sentry uses Seer autofix and Claude routines to build an automated debugging workflow that detects, fixes, and routes code issues automatically. ## Connecting OpenTelemetry Traces with Sentry Errors in One Trace Waterfall DevFeed: [Connecting OpenTelemetry Traces with Sentry Errors in One Trace Waterfall]() Original publisher: [Read original article]() Author: Johannes Daxböck; Neel Shah Published: 2026-08-05T09:00:00Z Content type: tutorial Language: en Sources: [Sentry Blog]() Topics: [OpenTelemetry](), [tracing](), [telemetry](), [SDKs]() Tags: [errors](), [instrumentation](), [logs](), [opentelemetry](), [sdk](), [sdks](), [sentry](), [tracing]() ### AI overview Sentry's OtlpIntegration connects OpenTelemetry traces with Sentry errors and other events, allowing them to appear together in a trace waterfall. It reads the active OTel trace context and associates it with Sentry events while configuring an OTLP exporter for Sentry. ### Source excerpt The OtlpIntegration bridges OTel traces and Sentry errors. Keep your OTel setup, add Sentry for errors, and see both in one trace waterfall. ## Sentry Unreal Engine SDK Adds Session Replay to Crash Reports DevFeed: [Sentry Unreal Engine SDK Adds Session Replay to Crash Reports]() Original publisher: [Read original article]() Author: Ivan Tustanivskyi Published: 2026-07-24T09:00:00Z Content type: release Language: en Sources: [Sentry Blog]() Topics: [session replay](), [Unreal Engine](), [SDK](), [debugging](), [tracing]() Tags: [crash](), [debugging](), [sdk](), [sentry](), [session-replay](), [trace](), [unreal]() ### AI overview Sentry's Unreal Engine SDK adds Session Replay, recording the last several seconds of gameplay and attaching the footage to crash reports. Developers can review rendered frames alongside context, breadcrumbs, logs, and traces to investigate what led to a crash. ### Source excerpt The Sentry Unreal Engine SDK now records the last seconds of gameplay and attaches the footage to crash reports, so you can watch what led to the crash. ## From 57 bugs to 1, thanks to Seer DevFeed: [From 57 bugs to 1, thanks to Seer]() Original publisher: [Read original article]() Author: Dan Mindru Published: 2026-07-23T09:00:00Z Content type: tutorial Language: en Sources: [Sentry Blog]() Topics: [bug](), [Automation](), [GitHub](), [pull-requests](), [AI-assisted coding](), [test]() Tags: [automation](), [bug-fixes](), [bugs](), [coding-agents](), [github](), [pull-request](), [unit-test]() ### AI overview A guest post explains how Sentry, Seer, and cloud coding agents can automate fixes for lower-priority bugs, generate pull requests, and help clear a backlog. ### Source excerpt Learn how to automate bug fixes using Sentry, Seer, and cloud coding agents to clear your backlog. ## .NET support for Godot is now generally available DevFeed: [.NET support for Godot is now generally available]() Original publisher: [Read original article]() Author: Serhii Snitsaruk Published: 2026-07-22T09:00:00Z Content type: release Language: en Sources: [Sentry Blog]() Topics: [Godot](), [C#](), [.NET](), [SDK]() Tags: [2-0-0](), [c](), [c-plus-plus](), [c-sharp](), [configuration](), [cpu](), [gpu](), [metrics](), [net](), [optimization](), [release](), [version-2-0-0]() ### AI overview Sentry's Godot Engine SDK 2.0.0 brings C#/.NET error reporting and application metrics to general availability. It captures managed-code errors alongside GDScript and engine issues, and supports metrics for performance and gameplay analysis. ### Source excerpt Sentry's Godot Engine SDK 2.0.0 brings C#/.NET error reporting and application metrics to general availability. ## How to structure a log DevFeed: [How to structure a log]() Original publisher: [Read original article]() Author: Kyle Tryon Published: 2026-07-21T09:00:00Z Content type: tutorial Language: en Sources: [Sentry Blog]() Topics: [Logging](), [Structured-data](), [TypeScript](), [ESLint](), [SIEM, Security, Observability]() Tags: [code](), [guide](), [lint](), [logs](), [programming](), [schema]() ### AI overview A practical guide to writing structured logs that are consistent, queryable, and useful for debugging in production. It covers event naming, flattened attributes, predictable low-cardinality values, severity levels, primitive-value constraints, and enforcing conventions with an ESLint plugin for TypeScript. ### Source excerpt A practical guide to writing structured, queryable, lint-enforceable logs with consistent event names, attributes, severity, and schema checks. ## MongoDB Query Tracing in .NET with Sentry + OTLP DevFeed: [MongoDB Query Tracing in .NET with Sentry + OTLP]() Original publisher: [Read original article]() Author: James Crosswell Published: 2026-07-20T09:00:00Z Content type: tutorial Language: en Sources: [Sentry Blog]() Topics: [MongoDB](), [OpenTelemetry](), [tracing](), [.NET](), [telemetry](), [SDK]() Tags: [instrumentation](), [integration](), [mongodb](), [net](), [opentelemetry](), [sdk](), [telemetry](), [traces](), [tracing]() ### AI overview This tutorial demonstrates how to trace MongoDB queries in a .NET application by combining MongoDB's built-in OpenTelemetry instrumentation with Sentry's OTLP ingestion support in the .NET SDK. It explains how MongoDB spans and query data can flow into Sentry for performance analysis and correlation with errors and replays. ### Source excerpt Trace MongoDB queries in .NET using OpenTelemetry and Sentry's OTLP ingestion. See spans, query data, and error correlation in Sentry. ## How Sentry replaced ASCII snapshots with structured context for Seer Agent DevFeed: [How Sentry replaced ASCII snapshots with structured context for Seer Agent]() Original publisher: [Read original article]() Author: Mihir Mavalankar Published: 2026-07-16T09:00:00Z Content type: article Language: en Sources: [Sentry Blog]() Topics: [context](), [dashboards](), [Document Object Model (DOM)](), [Claude]() Tags: [agent](), [claude](), [context](), [cost](), [history](), [structure]() ### AI overview Sentry describes replacing ASCII page snapshots with structured semantic context for its Seer Agent. The change is intended to reduce token costs and context degradation while enabling richer interaction hints and future page manipulation. ### Source excerpt How Sentry replaced ASCII page snapshots with structured semantic context to make Seer Agent faster, cheaper, and ready for agentic actions. ## Logging Best Practices for Debugging Applications DevFeed: [Logging Best Practices for Debugging Applications]() Original publisher: [Read original article]() Author: Ben Coe Published: 2026-07-09T09:00:00Z Content type: tutorial Language: en Sources: [Sentry Blog]() Topics: [Logging](), [debugging](), [observability](), [audit](), [Instrumentation]() Tags: [audit](), [best-practices](), [debugging](), [how-to](), [instrumentation](), [learn](), [logging](), [observability](), [sentry]() ### AI overview This article presents practical logging best practices for applications. It recommends adding targeted log lines, using temporary instrumentation during investigations, recording runtime decisions and intermediate outcomes, and auditing access and mutating operations to make production software easier to debug and understand. ### Source excerpt Learn logging best practices: what to log, how to structure messages, and what to avoid. Practical tips with Sentry's structured logging. ## How the Usage app uses Sentry to detect crashes, silent errors, and build regressions DevFeed: [How the Usage app uses Sentry to detect crashes, silent errors, and build regressions]() Original publisher: [Read original article]() Author: Dan Mindru; Oleh Stasula Published: 2026-07-02T09:00:00Z Content type: article Language: en Sources: [Sentry Blog]() Topics: [App](), [monitor](), [Processes](), [iphone](), [data]() Tags: [app](), [apple](), [dependency](), [errors](), [guest-post](), [iphone](), [mac](), [memory](), [monitor](), [network](), [process](), [sentry]() ### AI overview This guest post explains how the Usage system activity monitor uses Sentry to detect crashes, silent errors, build issues, and user feedback problems. The app runs across iPhone, iPad, and Mac and depends partly on undocumented APIs, so OS updates can change data structures without notice. Separate Sentry projects track the app's cooperating components and their failure modes. ### Source excerpt How the Usage app uses Sentry to catch crashes, silent errors, and build regressions before 1.6 million users notice. ## What an AI Agent Failure Revealed About Model Tradeoffs, Evals, and Traces DevFeed: [What an AI Agent Failure Revealed About Model Tradeoffs, Evals, and Traces]() Original publisher: [Read original article]() Author: Sergiy Dybskiy Published: 2026-07-01T09:00:00Z Content type: article Language: en Sources: [Sentry Blog]() Topics: [Traces](), [Artificial Intelligence](), [Prompt Engineering](), [vercel ai sdk](), [gateway]() Tags: [agent](), [ai](), [eval](), [evals](), [model](), [tool](), [tools](), [trace](), [traces]() ### AI overview An experiment with an AI-powered conference schedule builder showed that an agent could invent speakers, falsely claim to have retrieved them from an API, and appear grounded because a tool had run. The article discusses model tradeoffs, evaluation limitations, and the value of inspecting agent traces. ### Source excerpt I gave the free tier a cheaper model and it invented conference speakers who don't exist. What that taught me about model tradeoffs, evals, and reading agent traces. ## Next.js already traces your requests. Here's how to export them with OpenTelemetry. DevFeed: [Next.js already traces your requests. Here's how to export them with OpenTelemetry.]() Original publisher: [Read original article]() Author: Kyle Tryon Published: 2026-06-29T09:00:00Z Content type: tutorial Language: en Sources: [Sentry Blog]() Topics: [Next.js](), [OpenTelemetry](), [tracing](), [telemetry]() Tags: [next-js](), [opentelemetry](), [telemetry](), [traces](), [tracing](), [vercel]() ### AI overview This tutorial explains how to export the traces that Next.js generates automatically using OpenTelemetry and @vercel/otel. It covers sending traces to Sentry or other OTLP-compatible backends, understanding trace spans, and differences between Node and Edge runtimes. ### Source excerpt Next.js creates OpenTelemetry traces automatically. Learn how to export them to Sentry or any OTLP backend with @vercel/otel. ## Sentry's new AI grouping model reduces duplicate issues and incorrect merges DevFeed: [Sentry's new AI grouping model reduces duplicate issues and incorrect merges]() Original publisher: [Read original article]() Author: Kush Dubey; Yuval Mandelboum Published: 2026-06-12T09:00:00Z Content type: article Language: en Sources: [Sentry Blog]() Topics: [Artificial Intelligence](), [Machine learning](), [Embeddings](), [Algorithm](), [GPU]() Tags: [ai](), [algorithm](), [connection-pool](), [embeddings](), [gpu](), [issue](), [ml](), [model](), [production](), [server](), [timeout]() ### AI overview Sentry describes an upgraded AI model for grouping errors into issues. The model prevents 20% more duplicate issues and halves incorrect merges, separating errors with distinct root causes that the previous model combined. ### Source excerpt Sentry's new AI grouping model prevents 20% more duplicate issues while cutting incorrect merges in half. Here's how we trained and deployed it. ## Catch visual regressions with Snapshots, now in beta DevFeed: [Catch visual regressions with Snapshots, now in beta]() Original publisher: [Read original article]() Author: Max Topolsky Published: 2026-06-11T09:00:00Z Content type: release Language: en Sources: [Sentry Blog]() Topics: [Testing](), [ci](), [Front end](), [User experience (UX)](), [Test coverage]() Tags: [agent](), [ci](), [frontend](), [sentry](), [testing](), [ux]() ### AI overview Sentry Snapshots is in beta and provides screenshot diffing in CI to detect unintended visual changes on each commit. The article explains snapshot testing, its use as a guardrail for agent-generated changes, and how Sentry used Snapshots to improve its own toolbar UX. ### Source excerpt Sentry Snapshots is now in beta. Get screenshot diffing in CI to catch unintended visual changes on every commit, for any platform with a frontend. ## Works on my machine: how we use AI to reproduce reported bugs DevFeed: [Works on my machine: how we use AI to reproduce reported bugs]() Original publisher: [Read original article]() Author: Neel Shah Published: 2026-06-08T09:00:00Z Content type: article Language: en Sources: [Sentry Blog]() Topics: [Claude](), [SDKs](), [bug](), [Artificial Intelligence](), [Development](), [GitHub](), [Open Source](), [Python](), [Django](), [Boilerplate](), [Concurrency](), [legacy]() Tags: [ai](), [bugs](), [claude](), [concurrency](), [django](), [frameworks](), [github](), [legacy](), [open-source](), [python](), [sdks]() ### AI overview Sentry's SDK teams use a Claude skill to automate parts of reproducing reported bugs, including setting up boilerplate environments and playgrounds with the required parameters. The approach is intended to reduce triage time across the team's large SDK ecosystem. ### Source excerpt How Sentry's SDK team built a Claude skill to auto-reproduce bug reports, reducing triage time across 159 packages. ## Errors, traces, logs, metrics: when to reach for what DevFeed: [Errors, traces, logs, metrics: when to reach for what]() Original publisher: [Read original article]() Author: Sergiy Dybskiy Published: 2026-06-05T09:00:00Z Content type: tutorial Language: en Sources: [Sentry Blog]() Topics: [SIEM, Security, Observability](), [telemetry](), [tracing](), [Logging](), [debugging]() Tags: [debugging](), [errors](), [logs](), [metrics](), [sentry](), [telemetry](), [traces]() ### AI overview This developer guide explains when to use errors, traces, logs, and metrics during application instrumentation and debugging. It distinguishes the question each telemetry signal answers and describes how the signals support different workflows. ### Source excerpt Errors, traces, logs, and metrics overlap enough that it's hard to know which to use. Here's when to reach for each signal, with a real debugging walkthrough. ## How we cut build times by two-thirds by deleting our CMS DevFeed: [How we cut build times by two-thirds by deleting our CMS]() Original publisher: [Read original article]() Author: Eli Lennox Published: 2026-05-28T09:00:00Z Content type: article Language: en Sources: [Sentry Blog]() Topics: [Astro](), [Markdown](), [Claude Code](), [gatsby](), [Vite](), [React]() Tags: [astro](), [claude-code](), [code](), [development](), [markdown](), [vite]() ### AI overview Sentry describes replacing legacy Gatsby sites and a traditional headless CMS with Astro, Markdown, and Claude Code skills. The migration reduced build times from about 14 minutes to under 4 minutes and eliminated the cited CMS and marketing automation API failures. ### Source excerpt Sentry replaced its CMS with Astro, Markdown, and Claude Code skills -- cutting build times from 14 to under 4 minutes and eliminating API failures. ## You don't need to pick one: how Sentry and OpenTelemetry work together DevFeed: [You don't need to pick one: how Sentry and OpenTelemetry work together]() Original publisher: [Read original article]() Author: Lazar Nikolov Published: 2026-05-27T00:00:00Z Content type: tutorial Language: en Sources: [Sentry Blog]() Topics: [OpenTelemetry](), [telemetry](), [tracing](), [Traces](), [context](), [Instrumentation](), [React](), [SDKs]() Tags: [frontend](), [instrumentation](), [opentelemetry](), [sentry](), [telemetry](), [traces](), [tracing](), [w3c]() ### AI overview This article explains how to combine Sentry and OpenTelemetry in a web application. It recommends keeping OpenTelemetry for backend instrumentation, using the Sentry SDK on the frontend for browser context and application features, and exporting backend traces and logs to Sentry through OTLP directly or through an OpenTelemetry Collector. Sentry's OTLP ingest supports traces and logs, but not metrics at the time described. ### Source excerpt Use Sentry on the frontend, keep OpenTelemetry on the backend, and choose direct OTLP or Collector forwarding for OTLP events. [Next page]()