February 2018 Summaries

Internet Explorer and Edge throw errors in the end user's language, which can be challenging for developers who need to understand and resolve these issues. Sentry now automatically translates Edge and IE errors to English using a large indexed dictionary of almost every error across all supported languages. This helps group similar errors into a single issue, making it easier to measure an issue's impact and severity, and then appropriately resolve it once identified.

Exception Perceptions' is a new series that explores Observability through videos, tutorials, and technical examples. The series aims to provide answers to common questions about Observability best practices, such as using logs for debugging, understanding user behavior with errors, and integrating APM into the framework. The goal of the series is to help developers increase productivity while providing a comprehensive resource for all their Observability-related queries.

Matthew Machuga, lead analytics team at Auth0, recounts a recent deployment failure to a new AWS region due to Terraform's error message indicating an invalid number of nodes. The issue turned out to be a limitation in the specific AWS region, which was not yet available for that service. The team had to rewrite their infrastructure and made significant changes to their approach to avoid similar issues in the future, including increased investigation and planning ahead of time, tracking Amazon's release history, and using alternative technologies when necessary.

There are few things in life that we enjoy more than good, healthy, broken code. No one ever ships 100% perfect code.` This highlights the inevitability of bugs in software development and sets the stage for a discussion on how to improve debugging processes. The author suggests that instead of spending hours fixing a bug after it's occurred, developers can focus on preventing issues from arising in the first place. This can be achieved through better testing, including writing comprehensive tests to cover every scenario, as well as implementing robust QA processes that involve users as part of the testing team. However, even with thorough testing and quality assurance, bugs will still occur. In these cases, having a proactive approach to debugging is crucial. One effective method for capturing exceptions in real-time is by using global callbacks like `window.onerror` or its expanded version `error`. These callbacks provide valuable information about the error, including the line number and stack trace, which can be used to diagnose and fix issues quickly. Despite their potential benefits, these callbacks also have limitations, such as being affected by minification and CORS restrictions, which can make it difficult to obtain accurate error information.

Welcome Lauryn Brown Lauryn joins Sentry from Carnegie Mellon University in Pittsburgh, where she studied Information Systems and conducted research on using artificial intelligence to improve online learning for kids. As a researcher, she developed systems to help kids learn math and game theory. After college, she worked on a Lisp Interpreter built in COBOL at the Recurse Center. Now, she works on Sentry's Platform team to tackle challenging backend problems. Lauryn is excited to explore San Francisco as a new California resident.

Here's a 1-paragraph summary of the text, covering key points in an objective, neutral, and interesting tone: The Sentry Scouts Meetup event was a success, with attendees enjoying good snacks, entertaining comedy, and engaging discussions on open source projects. The event featured a panel discussion with notable speakers from various companies, including Honeycomb.io, GitHub, and Uber, sharing their experiences and tips on managing and contributing to open source projects. The meetup also included PowerPoint Karaoke, which was a hit among attendees. With its promise of fun, quality content, and no sales pitches or cold pizza, the event lived up to its hype, proving that developers indeed want to have fun while coding.

If you're using Sentry for error management in your projects, it's worth considering organizing them into multiple projects to improve manageability and relevance of issue streams for developers and customization of alert notification settings.

Vu Ngo has joined the Sentry team from Invoca, where he developed a standalone platform for integrating and managing integrations created by his team. In his free time, Vu enjoys checking the surf forecast and listening to Star Wars and Harry Potter audiobooks.

AdRoll, as a prospecting and retargeting platform, relies heavily on real-time monitoring of its system to handle over 70 billion requests daily, with each request needing to be processed in under 100ms. The company has a globally distributed infrastructure of up to 3,000 machines across the globe, generating seven trillion events every day, which are monitored for unexpected issues that could lead to costly errors and domino effects. AdRoll's monitoring philosophy prioritizes swift action over coddling instances, allowing engineers to approach issues calmly and utilizing a Blue-Green deployment strategy to minimize downtime and maximize efficiency.