Hi there, join our upcoming webinar introducing pganalyze Query Advisor and check out the recording from Lukas' talk at PGConf.dev on tracking plan statistics over time.
Welcome to the June 2025 edition of the pganalyze newsletter!
First up, we’re hosting a webinar introducing pganalyze Query Advisor, our new end-to-end query optimization workflow for Postgres. If you want to see how to automatically detect slow queries, get targeted recommendations, and track improvements over time, register today.
We also just made it into G2’s database monitoring market reports for the first time, picking up badges for Best ROI and Fastest Implementation. Thanks so much to everyone who shared reviews after our last newsletter!
Plus, if you caught us mentioning it last month, the recording from my PGConf.dev talk on our upcoming pg_stat_plans 2.0 extension is now available. Check it out to see how we’re building a lightweight, production-safe way to track plan changes over time.
And if you’re at RailsConf this week, come say hi to the team at our booth. Always happy to chat Postgres, swap stickers, or hear what you’re building.
As always, check out our product updates and list of upcoming events at the end of this email.
All the best,
Lukas
Introducing Query Advisor: Detect and Fix Slow Postgres Queries with pganalyze
Tuning slow queries is one of the most impactful ways to improve PostgreSQL performance, but finding and fixing inefficiencies is often manual, error-prone, and time-consuming. pganalyze Query Advisor changes that by introducing the first end-to-end query optimization workflow purpose-built for PostgreSQL, bringing automated detection of common query plan problems, actionable rewrite recommendations, and impact tracking together in a single tool.
Whether you’re a DBA managing hundreds of Postgres instances, a platform engineer looking to scale reliably, or a developer wanting to ship performant code, Query Advisor gives you a faster, more confident path to optimized queries.
PGConf.dev Recording: Tracking Plan Shapes Over Time with Plan IDs & pg_stat_plans
In case you missed it, we recently shared a behind-the-scenes look at a new open-source extension we’re building to capture per-plan statistics, which is similar to Aurora’s aurora_stat_plans, but designed to work anywhere you run Postgres.
This extension builds on pg_stat_statements to track when plans change and how different execution plans for the same query perform over time. Compared to auto_explain, which is great for catching outliers, this gives a complete picture of all the plans in use, with minimal overhead that makes it practical for production.
We talked through our early prototype at PGConf.Dev, and the recording is now available. Check it out to see how we’re shaping this extension ahead of the Postgres 18 release — and what it could mean for understanding your query performance at an entirely new level.
We asked in our last newsletter for your help to get us there, and you delivered. Huge thanks to everyone who shared reviews and feedback. It means a lot, and we’re always up for hearing more. Check out our interactive product tour to learn more about the features delivering ROI for your database.
We’re excited to be participating in several Postgres community events throughout 2025! We're wrapping up the week here in Philadelphia for RailsConf, and Ryan from our team will be speaking at PGDay Austria.
We’re looking forward to more great conversations and knowledge sharing at upcoming events across the Postgres community!
Amazon Aurora: Add support for Postgres 17 when aurora_stat_plans is enabled.
Previously, Amazon Aurora users on Postgres 17 that also had plan statistics enabled were unable to collect query statistics due to a column "blk_read_time" does not exist error.
Log Insights: Improve parsing Heroku auto_explain logs using JSON format.
Support newlines in the middle of the EXPLAIN query with the JSON format.
Add additional LDAP authentication settings for SSL certificate verification
Adds LDAP_ENCRYPTION_CA_FILE to set the path to CA file to validate server certificate (use a volume mount to provide this file to the container)
Adds LDAP_ENCRYPTION_VERIFY_MODE to set whether to verify the server certificate ("peer", the default) or not ("none")
Note that LDAP_ENCRYPTION_VERIFY_MODE=none restores the historic behavior of not verifying the server side certificate, as it was before v2024.10.0.
Update bundled collector from 0.65.0 to 0.66.2
Note that some of the new functionality (e.g. pg_stat_io tracking) is not yet enabled in this Enterprise Server release, as this update is mainly aimed to pull in collector-side bugfixes.
Improve reliability of partition creation for internal statistics tables
This replaces the mechanism that runs in the background to create daily partition for the internal database, with a more reliable method that runs directly in the statistics snapshot workers
Resolves errors like "no partition of relation ... found for row" that caused missing data in some cases
Improve LDAP self-test to accurately report SSL connection problems
Connection tracing: Don't show incorrect information from newer backends
Index Advisor: Don't fail with "internal error" in certain cases when handling queries involving 2 or more tables
Dependency security updates and other routine security updates to packages in the base image
