April Product Update
4 min read
•
4 days ago

Hello, April was a big month at Readyset. Let’s catch up on what’s new.
Introducing rdst: The Client-Side SQL Performance Toolkit
Our mission has always been to make database scaling as easy as possible. We’re excited to share the next step in that journey: rdst.
rdst is a production-native, AI-powered CLI that connects directly to your Postgres or MySQL database and helps you write better queries, diagnose performance issues, and close the loop on caching before bad queries ever reach production.
It analyzes your queries, execution plans, and runtime statistics to give you clear explanations and concrete next steps based on your actual workload. And it does it all in a few quick, simple commands, straight from your terminal.
Product Updates
Readyset Core
- MySQL GTID Support and Zero-Downtime Failover: Readyset now fully supports GTID-based replication for MySQL. Combined with new failover commands, you can redirect Readyset to a new upstream primary without restarting. The cache stays intact, queries keep serving, and replication resumes from where it left off. Read the full post.
- Query ID Label in Query Metrics: Readyset now includes a query ID label in recorded query metrics in the default non-verbose logging mode, making workload analysis easier.
- Runtime Failover: Introduced runtime failover commands that let you repoint to a new upstream primary and resume replication without restarting, enabling faster recovery and simplifying failover during planned or unplanned topology changes.
- MySQL Replication Batching, Up to 21x Throughput Improvement: MySQL replication now batches row events within a transaction into a single RPC. Initial tests showed up to 21x throughput improvement for large transactions. Controlled by the new
--replication-batch-size flag(default 50,000). - Group Commit, Up to 7.6x Throughput Improvement: Readyset now coalesces consecutive small transactions into a single RPC batch via group commit. Initial tests showed up to 7.6x throughput improvement for single-row autocommit workloads. Controlled by
--group-commit-max-trx(default: 20) and--group-commit-wait-us(default: 500). - More Queries Now Cacheable: AVG queries with
WHERE IN, range, orBETWEENconditions can now be cached, along withjson_object_aggandjsonb_object_aggqueries withWHERE INclauses. FLUSH ALL SHALLOW CACHES: New SQL command that clears cached data for all shallow caches without removing cache definitions.- Per-Query Execution Metrics via SQL: New virtual relations
readyset.upstream_query_statsandreadyset.cached_query_statslet you query per-query execution metrics directly via SQL.
Check out the full release here.
New on the Blog & YouTube
Replication Internals: Decoding the MySQL Binary Log (Parts 6–9)
Marcelo continues his deep-dive series into MySQL replication internals. Since the March update the series has reached Part 9, covering the Query Event, TABLE_MAP_EVENT, Row Events, and XID_EVENT. If you are building on or troubleshooting MySQL replication, catch up on this series:
- Part 6: QUERY_EVENT — DDL Statements and Transaction Boundaries
- Part 7: TABLE_MAP_EVENT — Table Metadata for Row-Based Replication
- Part 8: Row Events — WRITE_ROWS, UPDATE_ROWS, and DELETE_ROWS
- Part 9: XID_EVENT — Transaction Commit
Readyset Is Ready for MySQL 9.7 — Three Commands with rdst: MySQL 9.7.0 LTS launched and Readyset was ready with support from day one. Three rdst commands take you from nothing installed to sub-2ms cached reads off a live 9.7 instance.
Query Hints: Let Your Application Decide What Gets Cached: With Readyset Query Hints, add a single SQL comment and create a cache on first execution. No DDL, no DBA, no separate deployment.
Why LLMs Write Incorrect SQL (and What That Means for Your Database): Most LLM-generated SQL runs, returns results, and passes code review. This post breaks down why that's exactly what makes it dangerous, and what teams are doing to catch the problems before they become incidents.
Inside MySQL: 20 Years of Source Code, Open Source Contributions, and What Comes Next: Readyset engineer Marcelo Altmann talks through 20 years inside the MySQL codebase, what it takes to contribute upstream, and where MySQL is headed next.
April Event Recap:
Introducing Readyset Shallow Cache Webinar
Earlier this month we hosted a live walkthrough of Readyset Shallow Cache, showing how Shallow Cache and Deep Cache work together to reduce database load across your full workload with zero ramp-up time. If you missed it, the recording is now live.
🗓️ Upcoming Events
rdst Fleet Audit: Live Demo: May 19, 8:00 AM PT / 11:00 AM ET
We are hosting our next live webinar in May to introduce rdst. We’ll run rdst against a real workload, break down what is driving load, and surface top bottlenecks and optimization priorities. We’ll also show how rdst guides you through rewriting inefficient queries, eliminating unnecessary replicas, and deploying caching directly from your terminal.
No guesswork, no extra infrastructure. If your database costs are climbing and you are adding replicas without knowing what’s driving the load, this one is for you. Register here.
Vancouver Web Summit - May 11-14, Vancouver BC
We’ll be in BC next month for Vancouver Web Summit. If you’re attending or nearby, book time with our team to see how to scale your database performance and save on cost without adding infrastructure.
Percona Live 2026 - May 27-29, Mountain View, CA
We’re looking forward to sponsoring Percona Live next month! Catch Vinicius’s talk “Beyond Manual Caching: How Readyset Offloads MySQL and Postgres Workloads” on May 28. Stop by our stand if you want to learn more.
🧑💻Request Features
Keen to use Readyset but your database or cloud environment is not supported? Reach out to [email protected] and let us know.
Get Involved!
Questions, bugs, ideas, or just want to say hi? You can find us on GitHub, Slack, and LinkedIn. Looking to contribute? Check out our newcomer issues, it's a great place to start.
Authors