May 2026 Summaries

The blog post by Simeon Griggs focuses on optimizing database queries to reduce egress, which is the data transferred out from a database over the public internet, often billed by cloud providers. It highlights the importance of making query responses smaller and less frequent to enhance application performance and reduce costs. Strategies include selecting only necessary columns, using pagination, implementing caching, and considering private connections like AWS PrivateLink to avoid public internet data transfer. The text also details the use of Postgres features, such as jsonb_agg() for JSONB column filtering, and tools like PlanetScale Insights to monitor query sizes and egress usage patterns. It emphasizes the proactive optimization of database queries to achieve faster responses, cleaner code, and lower resource demands, encouraging developers to audit their code for unnecessary data requests and to leverage query tagging for better performance tracking.

PlanetScale Insights is a tool designed to help users identify and resolve database performance issues by providing detailed metrics on query performance, latency, and resource usage. It supports PlanetScale Postgres and Vitess/MySQL databases, offering features such as latency percentile tracking, query grouping with tags, and traffic control to limit resource consumption. Users can investigate individual queries, identify slow or inefficient ones, and receive recommendations for improvements, such as adding indexes. The tool also highlights anomalies and errors, allowing for proactive database management and optimization. By using PlanetScale Insights, database administrators can enhance application performance by addressing problematic queries and improving schema design, ultimately leading to more efficient and reliable database operations.

Database benchmarks, though imperfect, are valuable tools that help users understand tradeoffs and make informed decisions about infrastructure. However, many cloud vendors have restrictions known as "DeWitt clauses" that limit comparative benchmarking, a practice counter to the legacy of database researcher David DeWitt, who advanced the industry by publishing real system measurements. PlanetScale is moving away from such restrictions by updating their Acceptable Use Policy (AUP) to allow and encourage fair and transparent benchmarking, provided it adheres to best practices. They stress that benchmarks should be conducted in good faith, avoiding misrepresentation or interference, and they advocate for a new standard that emphasizes depth, technical soundness, and reproducibility. PlanetScale invites other vendors to adopt similar standards in their AUPs, encouraging transparency and allowing customers to effectively compare systems.

Benchmarking databases is a complex process that requires careful consideration of several factors to ensure accurate and unbiased results. Key principles include understanding the client-server architecture, where latency between the benchmark client and database server can significantly impact results. Resource allocation must be balanced to prevent misleading comparisons, and workload characteristics such as traffic ratios and query patterns need to be carefully configured. The choice between open and closed-loop benchmarks is essential, as each serves different purposes: open loops simulate real-world client demands, while closed loops offer controlled comparisons. Throughput and latency are critical metrics, with percentiles providing a more comprehensive view of latency than averages alone. Proper benchmarking involves warming up caches, aligning database configurations, and accounting for environmental inconsistencies, especially in cloud environments. To ensure meaningful comparisons, benchmarks should compare similar technologies and document setups thoroughly for reproducibility. Common pitfalls include failing to measure at steady state, overlooking client bottlenecks, and reporting only averages without context. By adhering to these practices, benchmarks can provide valuable insights into database performance.