October 2020 Summaries

The Sentry team is making changes to increase the reliability of their service, which may require customer-side updates to firewall rules. The IP addresses associated with Sentry will be changing on December 1st, 2020, and organizations will need to update their applications to use a new organization subdomain IP address. This change aims to improve latency for applications sending events from outside North America by utilizing an event ingestion service behind an Anycast address. Organizations can find their subdomain by visiting the Sentry settings page and should reach out to support if they have any questions about the impact of this change on their service.

Sentry is a powerful tool for capturing runtime errors and has recently been enhanced with features such as performance monitoring and release health. In the context of code refactoring, Sentry can be particularly useful in identifying deprecated code paths that are still in use. By leveraging Sentry's capabilities, developers can track down remaining call sites and fix them before removing the old code paths, ensuring a smooth transition during refactoring projects. The author of this case study used Sentry to update icons in their application, replacing deprecated forms with new React components, and successfully tracked down and fixed all remaining call sites using Sentry's event tracking capabilities.

Sentry has launched a major update to its Java SDK, which includes automatic error grouping, alerting, breadcrumbs, and more, specifically designed for Spring Boot applications. The new integration with Logback allows for seamless forwarding of error logs to Sentry, while also collecting breadcrumb events with log levels lower than error. To configure Sentry in a Spring Boot application, developers can add the `sentry-spring-boot-starter` dependency and set the `sentry.dsn` property, which enables real-time reporting of unhandled exceptions and detailed metadata. The integration also supports attaching release information, such as Git commit IDs, and user data, including cookies and IP addresses, with optional configuration. Additionally, Sentry integrates with other frameworks like Log4j2 and traditional non-Spring servlet applications.

Debug errors in Lambda functions often require combing through logs and searching for relevant terms, but using Sentry can provide detailed error stack traces and debugging context. To instrument a Node function with Sentry, developers need to create a dummy function, add the Sentry package to their Node folder, initialize Sentry in the Lambda function, and wrap their handler. This provides visibility into errors from Lambda functions, including detailed stack traces, function context, and deep links to CloudWatch logs. With customizations such as capturing timeout warnings and adding breadcrumbs, developers can quickly identify issues, reproduce them in their local environment, and get to the root cause of problems.