March 2026 Summaries

Logging in Next.js can be challenging due to its ability to run in three different environments: Node.js, Edge, and the browser, with most existing logging tools primarily designed for Node.js. This results in incomplete logging, especially for client-side and Edge environments. To address these challenges, it is essential to use logging solutions like LogTape or Sentry, which are runtime-agnostic and can capture logs across all environments. Additionally, incorporating trace-connected structured logging can significantly enhance debugging by associating logs with unique trace IDs, allowing for comprehensive querying and aggregation. This setup facilitates identifying and resolving issues by providing a complete picture of the application's behavior across different runtimes. Both LogTape and Sentry offer ways to manage structured logs effectively, with Sentry offering built-in logging capabilities as part of its SDK, making it a suitable choice for those already using Sentry for error tracking and tracing. Implementing a logging strategy that captures logs from all parts of a Next.js app and utilizes structured data can greatly improve the observability and maintainability of applications.

The blog post discusses the observability gaps in Next.js applications and provides guidance on how to address them using Sentry. Next.js offers features like server-side rendering and file-based routing but lacks visibility into production environments due to its three-runtime architecture (client, server, edge). This results in errors that can be challenging to trace, particularly when database queries are masked by ORM abstractions, and server actions lack detailed error messages. The post highlights specific issues such as hydration errors in React, which occur when client-rendered HTML does not match the server-rendered content, and the lack of telemetry spans in server actions. To improve observability, the blog suggests using Sentry to capture full stack traces, provide HTML diffs for hydration errors, manually instrument server actions, and integrate database clients to expose SQL queries within traces. It also covers AI agent monitoring using the Vercel AI SDK and emphasizes the importance of configuring logs and metrics properly, as these are not subject to sampling like traces. The post concludes by encouraging the use of Sentry's tools to gain a comprehensive view of application performance across all Next.js runtimes.

Seer, an AI agent used by Sentry to identify bugs and suggest fixes, played a crucial role in diagnosing and resolving an outage that it inadvertently caused. On February 21, 2026, Seer's AI-powered issue summarization failed in the EU region due to an upstream issue with Google's Vertex AI platform, compounded by a flaw in Seer's own code that incorrectly blocklisted its provisioned throughput (PT) region. This led to a cascade of blocklisting across all EU regions, resulting in a total outage. The root cause was a missing code exception for the PT region in the blocklist logic, which was present in the US deployment but not in the EU deployment. Seer was instrumental in quickly identifying the issue by analyzing error patterns, which helped engineers realize the oversight and implement a six-line code fix to prevent PT regions from being blocklisted, ultimately restoring service. The incident highlighted the critical importance of ensuring that application code is fully aware of infrastructure provisioning to avoid turning minor upstream issues into major outages.

The Sentry SDK has evolved from a simple crash reporter to a comprehensive observability tool that includes features like session replay, structured logs, AI monitoring, and automatic OpenTelemetry tracing. Despite these advancements, many users still operate on outdated versions, missing out on significant functionalities and improvements. The SDK's current iteration provides enhanced capabilities such as linking session replays directly to errors and traces, enabling a deeper understanding of application issues. Additionally, the transition to newer versions offers performance improvements and crucial security patches. The upgrade process, while potentially challenging due to API changes, is supported by detailed migration guides and AI-enabled tools that automate parts of the process, helping users efficiently transition to the latest version and fully leverage the SDK's capabilities.

In the age of AI, generative technologies are significantly impacting software licensing, challenging traditional frameworks like Open Source (OSS) and introducing new paradigms such as Fair Source (FSS). The rise of AI models, such as Opus 4.5 and Codex 5.2, has enabled AI to function as standalone agents, increasing reliance on AI-generated code and raising questions about copyright protection, as AI outputs may not meet the human creativity requirements for copyright. Amidst this shift, Sentry, a proponent of Fair Source, argues that FSS remains a viable licensing option as it provides enforcement through contractual obligations rather than relying solely on copyright infringement claims. This approach helps navigate the complexities introduced by AI, as FSS emphasizes non-compete clauses and industry collaboration without compromising business models, distinguishing itself from traditional OSS in an evolving digital landscape.

Choosing the right JavaScript logging library is crucial in modern development workflows, especially with AI playing a larger role in coding. Logging libraries offer several advantages over traditional console logging, such as improved performance, structured outputs, and the ability to filter and redact sensitive data. Pino is recommended for Node backends due to its speed and small bundle size, while Winston is preferred for its extensive configuration options and transport capabilities. Bunyan is noted for its simplicity but is not advised for new projects, and LogTape stands out for its universal runtime compatibility across various platforms. Sentry, although not solely a logging library, provides robust logging features integrated with its monitoring platform. Each library supports custom transports, allowing logs to be directed to various backends, enhancing debugging and correlating issues.

The text provides a comprehensive guide on routing OpenTelemetry (OTLP) logs to Sentry, outlining the benefits and setup process for integrating these two systems. It highlights the advantage of using OTLP over the native Sentry SDK, particularly for those who have already implemented OpenTelemetry and require vendor-neutral instrumentation, or need to send logs to multiple backends. The guide details how to configure a sample payment processing app to direct logs to Sentry's OTLP endpoint, explaining the setup of environment variables and the necessary credentials. It also contrasts the OTLP approach with the native Sentry SDK, noting the latter's automatic features like issue creation, session replay, and error correlation, which are not yet fully available in the OTLP beta. Furthermore, the text emphasizes adding rich attributes for easier debugging and discusses troubleshooting common issues, such as missing logs or high memory usage, providing solutions and best practices for effective log management.