May 2025 Summaries

The text introduces Session Health in Sentry, a feature that provides a clear view of stability and user experience. It allows developers to drill into each dip in session health with the connected issues, enabling them to push quick fixes or roll out releases with confidence. The feature offers various metrics such as crash-free sessions, new issues by release, total sessions by release, and more, providing direct visual connections between metric data and code changes. By using Session Health, developers can quickly pinpoint anomalies, analyze root causes, and prioritize fixes for affected users, ultimately deploying with confidence and ensuring a better user experience.

The article describes a slow query that led to a cascading failure in the `placid.app` SaaS application, which generates creatives at scale. The query calculated a user's credit balance and was triggered by API endpoints handling creative generation, causing thousands of concurrent slow queries due to high throughput requests. The developer identified the issue by analyzing monitoring data and error logs with tools like Sentry, and fixed it within 10 minutes. To prevent similar issues in the future, the application has optimized the query by storing partial credit balance aggregations and implemented fallback strategies to prevent repeated database hits after failures.

The Sentry Cocoa SDK has improved app hang detection for iOS and tvOS, providing better visibility into how, when, and why apps hang. This update gives developers deeper insights into the kinds of hangs that impact users and the context needed to prioritize and fix them. The new functionality detects fatal app hangs, differentiates between fully blocking and non-fully blocking app hangs, measures duration for non-fatal app hangs, and tracks app hang rate. With this improvement, developers can better understand how many of their users are affected by fatal app hangs and make data-driven decisions to resolve issues faster.

The text highlights the importance of context in debugging software issues, particularly when using AI-powered tools. The current state of AI models lacks this critical component, making it challenging for them to accurately identify and fix errors. Sentry, a popular tool for error tracking, provides comprehensive context that enables developers to debug issues more efficiently. A new AI agent called Seer is being developed to leverage this context and provide even better debugging capabilities, including identifying root causes 90% of the time, generating code fixes, and opening pull requests. By prioritizing context, Seer aims to enhance the debugging process, making software development more efficient and reliable.

Sentry is now in open beta, offering a feature to send application logs directly to Sentry, providing context for debugging issues. This integration allows developers to view logs alongside errors and traces, improving the debugging workflow. With structured logs, users can search by metadata such as customer ID, payment ID, and trace ID, enabling them to quickly identify root causes of issues. The feature is available in under a minute for Python, JavaScript, Go, Ruby, Java, and Android, with more SDKs to be supported soon. Sentry aims to help developers find and fix issues faster by providing better context at the moment they're dealing with an issue.

Monitoring your MCP Server in Production (with Sentry)` provides an overview of how to monitor an MCP server in production using Sentry for errors, traces, and logs. The guide covers instrumenting the Cloudflare Workers and Durable Object with Sentry, setting up error handling, tracing, and logging, as well as configuring Sentry for stdio mode. It also discusses abstracting context and provides examples of how to use Sentry's features effectively. Overall, the guide aims to help developers set up a robust monitoring system for their MCP server using Sentry.

With Sentry's new Query Engine, tracing data can be transformed into actionable insights to address complex performance and reliability issues. The new Trace Explorer and Span Metrics feature allows users to query their raw data to spot outliers, identify patterns, create alerts, and root cause problems. Users can filter and group spans by various attributes, such as operation, service name, or custom attributes, to gain a deeper understanding of their application's performance. This enables proactive use of tracing, reducing the need for manual debugging and optimizing performance. With this new feature, users can visualize broad-stroke trends and spikes in metrics, create dashboards and alerts, and even attach custom attributes to spans to unlock more insights.

Sentry helped Amanotes identify the root cause of their Unity game's Application Not Responding (ANR) errors, which were not apparent through traditional profiling tools. By analyzing Sentry data, they discovered that scene loading transactions and heavy methods were causing delays, particularly when initialized too early during startup. Optimizing these areas led to a significant reduction in ANR rates by 50%, from 0.98% to 0.46%. The key improvements included reducing scenes' loading time, minimizing dependencies, preloading assets smartly, and deferring instantiation. By using Sentry for Unity, Amanotes was able to pinpoint invisible performance issues and optimize their game's startup and scene loading processes, resulting in a better user experience.

Sentry's new Laravel Insights provides a clearer picture of what's happening inside an app by surfacing errors, slow jobs, database queries, cache usage, and route performance, tying it all back to the issues being already tracked and how they affect users. This tool gives developers a solid starting point while allowing them to go as deep as they want without fighting the tooling. With full-stack distributed tracing, insights can be connected across different monitoring tools and platforms, including frontend and backend applications built with various technologies. By automating connections between these components, Laravel Insights helps developers identify performance issues before customers notice, making it easier to fix problems before they become major issues.

Today I'm thrilled to announce that Emerge Tools is joining Sentry, a company well known for universal and best-in-class crash reporting and application monitoring. Mobile has always been a complicated story in our industry, with many vendors neglecting it. At Sentry, we view mobile applications as an important extension of your business, deserving the same attention as the rest of your technology stack. The addition of Emerge Tools will bring new functionality to mobile teams, allowing them to ship smaller and lighter apps, remove dead code, manage internal distributions, and use visual regression testing. With this integration, Sentry aims to strengthen its existing products and extend its support beyond production monitoring, focusing on preventing bugs rather than just catching them.

When a checkout flow breaks, customers disappear quickly. To set up observability for critical paths like customer checkout, Sentry can be used with minimal effort. Setting up monitoring and distributed tracing was straightforward, allowing the capture of performance metrics, errors, and tracing data across both Flask and React applications. A dashboard was created to monitor the most critical aspect of an eCommerce storefront: the checkout flow. The instrumentation in the Sentry SDK allowed for the attachment of attributes to spans, such as the number of items added to a cart, enabling the tracking of key signals in users' most critical journeys. This improved the ability to spot "outages" that aren't just uptime or spikes in errors. In this case, a complete dropoff in customers adding items to their carts was detected, and drilling into errors revealed issues with Flask's backend and React's frontend, which were resolved quickly. Sentry allows for the creation of alerts on span metrics, routing them out to the right team, providing clarity without context switching chaos.

When a checkout flow breaks, customers disappear quickly. However, setting up observability with Sentry makes it painless to track user interactions and fix issues. By turning on Distributed Tracing in both frontend and backend applications, they captured performance metrics, errors, and tracing data. A dashboard was created to monitor the critical aspect of their eCommerce storefront: the checkout flow. The instrumentation allowed them to attach attributes to spans, such as the number of items added to a cart. This enabled them to spot issues like a complete dropoff in items added to carts, which they could drill into and see the root cause. They fixed a classic coding slip-up by initializing an empty array instead of None for product inventory. To improve their approach, they set alerts on span metrics and routed issues to the right team, making observability without context switching chaos possible with just a few lines of code.