Planet for the Valkey Community

Amazon ElastiCache for Valkey now features ElastiCache Insights, a pre-built dashboard in Amazon CloudWatch for your node-based clusters. ElastiCache Insights brings cluster health, memory, compute, network, command throughput, latency, keyspace, and errors into a single view sourced from a set of OpenTelemetry metrics that ElastiCache publishes automatically.

You can use ElastiCache Ins [...]

Valkey is a growing project, launched as a fork of Redis over two years ago, supported by twelve maintainers and nine committee members from various companies. Recent developments include enhanced throughput, improved cluster functionality, and plans for features like synchronous replication and tiered storage. The project also emphasises security and release efficiency through automation.

When Redis changed its licensing in 2024, replacing its permissive BSD license with proprietary, source-available alternatives, it prompted a major

The post Redis by proxy: Percona targets the last major hurdle to Valkey adoption appeared first on The New Stack.

Some open-source projects are hesitant to talk about how they use AI to improve their code. Not Valkey. Its developers are proud AI users.  Prague — Valkey, in case you don't know, is an open-source, high-performance, in-memory key-value data store used as a database, cache, message broker, and streaming engine. It's a community-governed fork of [...]

The post Valkey 9.2 Targets Memory O [...]

Protect live-event telemetry ingestion with Valkey Router: batch small beacons, group related events and buffer bursts before downstream processing.

I’ve encountered DBAs and ops engineers who shared war stories of running production long before automation and modern tooling went mainstream. Experiencing sudden primary faults during sleeping hours, getting paged, logging in to squint over trace logs that show replication lag on all replicas and working out which one is furthest ahead, promoting it, editing […]

The post Aut [...]

Amazon ElastiCache for Valkey now publishes OpenTelemetry metrics to Amazon CloudWatch for your node-based clusters, and each metric carries attributes you can filter and aggregate on with Prometheus Query Language (PromQL) expressions. ElastiCache now offers two monitoring modes. Standard monitoring, the existing experience, continues to publish CloudWatch Metrics (Classic) every 60 seconds and [...]

Reuse KV cache across inference workers with a shared Valkey cache, RDMA/EFA networking, plus NVMe and object storage tiers, all powered by Momento.

Introduction This RAM/Memory reduction is most useful for Valkey 9.1 environments that are key string heavy. They are used extensively in caching, sessions/authentication, real-time apps, ecomm, Ad tech/martech, and gaming. These scenarios see a reduction of 37.5%, which has been attained in our testing environments versus version 7.2.. The biggest jump in the reduction came … Continued

Amazon ElastiCache Serverless for Valkey now supports public endpoints, letting you connect to your cache directly from a laptop, a serverless function, or any application running outside AWS, without setting up a VPN, a bastion host, or an SSH tunnel.

With a public endpoint, ElastiCache Serverless gives you a fully managed cache reachable over the internet, with no VPC to configure and n [...]

Imagine this: your application is running fine, until one day, out of the blue, requests start timing out. The only thing you know for certain is that you implemented Valkey to be somewhere in the request path. Is it memory pressure? A lagging replica? A burst of slow commands? Without metrics, "somewhere in the request path" is as specific as your diagnosis gets.

Enter Prometheus. This p [...]

A 50-million-member sorted set benchmark compares memory use and insert throughput across six Redis releases and five Valkey releases.

Valkey rarely runs on its own. It often sits in front of a database, and when a page gets slow the answer is either a cache that is not being hit or a query that got worse. Checking both usually means one terminal on valkey-cli and another on the database, and holding the two halves of the picture in your head.

That is the situation I want to walk through, using a graphical u [...]

With Aiven Runtime you can deploy and run applications from source code in a GitHub repository directly in your Aiven projects. Host your applications right where you data is, and seamlessly integrate with Aiven for Apache Kafka®, Aiven for PostgreSQL®, Aiven for OpenSearch®, and Aiven for Valkey™ services. The Aiven Platform automatically handles the underlying networking, and container orchest [...]

This release brings MySQL-style grants to more service user resources, and expands OpenTelemetry and Valkey configuration options.

Highlights for this release:

• Aiven for Kafka, MySQL, OpenSearch, PostgreSQL, and Valkey user resources and data sources now expose a mysql_grants field for managing MySQL grants on the service user.
• New OpenTelemetry integrat [...]

Everything you need to monitor Valkey is already in the Prometheus ecosystem. This post is about how kube-prometheus, redis_exporter, and BetterDB Monitor come together to give you a complete picture of a running Valkey cluster, one tool per observability dimension.

Here's the one-minute version.

A Valkey cluster has two distinct observability dimensions: what's happening ins [...]

This release adds automated credential rotation, along with several smaller reliability and configuration improvements.

You no longer need to rotate service user passwords manually. With the new

ServiceUserRotation kind, you can manage a rotating pool of service users behind a stable connection Secret. Credentials can be rotated without disrupting dependent workloads.

When two questions need the same answer, generating it twice adds cost and delay. A semantic response cache lets an AI assistant reuse an earlier answer even when the next person phrases the question differently.

We published five security advisories in Valkey's first 21 months, from the March 2024 fork through the end of 2025. We published seven advisories in the first eight months of 2026 and shipped almost three dozen security-adjacent fixes. In one two-week stretch, three researchers who did not know about each other reported the same bug: a use-after-free in the script debugger, an uncommonly used f [...]

The KV cache is huge, so shipping it over a network should be a non-starter. I benchmarked DRAM, local NVMe, and remote Valkey to find out what the hop really costs, and the answer wasn't what I expected.

Nobody goes looking for a big key until something else breaks. One shard hits its memory limit while the others sit half idle, or p99 climbs with no change in traffic, and somewhere in the keyspace is a multi-megabyte hash nobody remembers creating. Valkey Admin 1.1 adds big key detection, which ranks the largest keys across every primary of a running cluster in one view, so you can find it with [...]

Momento Cache puts high-performance Valkey at your fingertips. Cluster provides direct control over topology, while Flex automatically optimizes resources.

This release adds new configuration options across several services and promotes beta VPC peering resources to general availability.

Highlights for this release:

• Aiven for Kafka, MySQL, OpenSearch, PostgreSQL, and Valkey user resources now expose a password_encryption_type field showing the hashing algorithm used for the stored password.
• Aiven for Kafka [...]

Valkey 9.0 introduced Atomic Slot Migration (ASM), a new resharding mechanism that replaces the traditional key-by-key migration process with a replication-inspired approach. Instead of copying individual keys using MIGRATE, the source node establishes a dedicated connection to the target node and streams the contents of the migrating hash slots, followed by incremental updates unti [...]

Fifty extra indexes don't slow down your writes, but one vector field slows them down 11x. I wanted to figure out why valkey-search behaves this way, and it wasn't what I expected.

The role of databases has shifted significantly as modern applications must deliver real-time analytics, dashboards, and machine learning alongside low-latency transaction processing. Handling these diverse demands with a single relational database has become unsustainable under growing data volumes. Consequently, organizations are adopting specialized database architectures where multiple engin [...]

Benchmarking your Valkey deployment allows you to understand how your infrastructure behaves under load and where its limits are. And Valkey leverages this with an easy-to-use tool that allows you to get actionable figures for your deployment and to evaluate that performance.

This blog post explores the functions of the tool and some common traps you might encounter on your first run.

A filter that looks like a query detail can change how Valkey searches your vectors. By the time you notice, the important decisions are already behind you.

Valkey 9.1 introduces three optimizations that reduce the per-key memory overhead without requiring any new commands or configuration changes.

These reduce the memory required to store strings and sorted sets, up to 44% overhead for string keys and up to 8.5 bytes per sorted set member for typical short members. These savings add up quickly when millions of keys are in use.

This bl [...]

Amazon ElastiCache now supports Graviton4-based M8g, R8g, and C8gn node families for Valkey and Memcached. Graviton4-based nodes provide up to 47% higher throughput, up to 43% lower P99 latency, and up to 31% better price-performance for on-demand pricing over Graviton3-based nodes of equivalent sizes on Amazon ElastiCache for Valkey, depending on node family, size, and workload configuration.

Dumanshu Goyal discusses optimizing data layers for low-latency workloads like AI feature stores. Drawing lessons from NASA's Space Shuttle, he explains how proxy architectures introduce hidden CPU costs, elevated tail latencies, and blast-radius risks. He demonstrates how direct-access Valkey architectures achieve microsecond latency, improve resilience, and slash infrastructure costs.

ClusterControl 2.5.0 is here, and it marks a milestone for the platform: ClickHouse joins the family of supported database engines — and with it, a capability no analytics vendor’s cloud can offer you. Wherever you run ClusterControl — on-premises, in any cloud, or hybrid — you can now deploy ClickHouse either as a standalone OLAP […]

The post ClusterControl 2.5.0 brings Click [...]

Modern AI inference stacks have split into layers, and each layer needs memory-speed access to something. Semantic caching needs vector similarity over recent prompts. Inference engines need a place to park KV tensors between requests. Retrieval-augmented generation (RAG) pipelines need filtered retrieval in a single round trip. The layer in front of all of it needs counters and rate limits that [...]

The community gathered at the Valkey Unlocked Conference to celebrate two years of innovation and ecosystem growth. In a keynote titled, "The Power of Community," Sailesh Krishnamurthy (VP of Engineering at Google) shared core insights on what it takes to build a thriving, resilient open source project and where the technology goes from here.

With Harbor v2.15.2, the Harbor team replaced Redis with Valkey as its internal cache backend. This was a deliberate open source choice that highlights the flexibility and resilience of the open source ecosystem.

Harbor is a CNCF graduated open source container registry, a place where organizations can store, manage, and secure container images, Helm charts, OCI artifacts, CNAB (Cloud Nat [...]

Across four releases, I/O-threading changes removed a serial copy bottleneck, cut p99 latency, and brought large GETs to line rate on our 200 Gbps test rig.

If you’ve used Valkey for very long you’re probably aware of the different primary data types in the core: strings, hashes, lists, sets, sorted sets, and streams. Each data type has an accompanying set of commands that manipulate that data type. For example, you use HSET to set a field and value pair to a hash, but you don’t use, for example, HSET to add an element to a list. What i [...]

DoorDash has developed Entity Cache, a transparent proxy caching platform built on Envoy and Valkey to reduce redundant service-to-service requests across its microservices architecture. Operating within DoorDash’s service mesh, the platform serves over 1.5M requests per second with 99.99999% availability through caching, event-driven invalidation, failure handling, and performance optimization [...]