Redis: Deep Dive into Performance Insights

Transcript

Hey there, Redis enthusiasts! Welcome back to another episode. I'm your host, and wow, do we have some fantastic improvements to dive into today. It's March 23rd, 2026, and the Redis team has been absolutely crushing it with some really thoughtful enhancements that are going to make our lives as developers so much easier.

You know what I love about today's changes? They're the kind of improvements that show the Redis team really understands what we need in production environments. These aren't just flashy new features - they're the solid, foundational work that makes Redis even more reliable and observable.

Let's start with something that's going to be a game-changer for anyone who's ever had to debug slow performance issues. Mincho Paskalev has added comprehensive slowlog metrics that give us way more insight into what's happening under the hood. Now, when you run INFO STATS, you're going to see three new metrics that tell the whole story of your slow commands.

First, there's slowlog_commands_count - this tracks every single command that's been written to the slowlog, even the ones that got trimmed out. Then you've got slowlog_commands_time_ms_sum, which adds up all those execution times, and slowlog_commands_time_ms_max for finding those really problematic outliers. But here's where it gets even better - INFO COMMANDSTATS now breaks this down per command type. So you can see exactly which commands are causing slowdowns and how often.

This is the kind of feature that's going to save so many late-night debugging sessions. Instead of trying to piece together performance issues from scattered logs, you've got everything right there in your Redis info output. It's like having a performance dashboard built right into your database.

Now, the second major change today is a bit more technical, but it's equally important. Moti Cohen has done some serious architectural work on the KEYSIZES tracking and ASM background operations. This is one of those commits that makes me really appreciate the engineering that goes into Redis - it's a complex refactor that touches a lot of systems but makes everything more consistent and reliable.

The key improvement here is around how Redis handles background operations, especially during ASM trimming. Previously, there could be consistency issues when the system was trying to update histogram data while background operations were running. Moti introduced this concept of an ASM trim context that keeps everything coordinated, so your diagnostic data stays accurate even when Redis is doing heavy lifting in the background.

What I find really interesting about this change is how it shows the evolution of Redis from a simple key-value store to this sophisticated system that needs to handle complex concurrent operations while maintaining data integrity. It's not the sexiest work, but it's absolutely essential for production reliability.

Both of these changes reflect something I really admire about the Redis development philosophy - they're constantly improving the observability and reliability of the system. These aren't breaking changes or dramatic rewrites. They're thoughtful enhancements that make Redis better at telling you what it's doing and more consistent in how it does it.

For today's focus, if you're running Redis in production, I'd encourage you to start thinking about how you can leverage these new slowlog metrics. They're going to be incredibly valuable for setting up alerts and understanding your application's performance patterns. Take some time to explore what INFO STATS and INFO COMMANDSTATS show you now - you might discover performance insights you never had access to before.

That's a wrap for today's episode. Keep building amazing things, and remember - every commit, every improvement, every thoughtful refactor is making our tools better. Until next time, happy coding!