Valkey

Why give a Valkey key a really long TTL, like 3 years?

A long TTL keeps cached data around without expiring anytime soon. But with a volatile eviction policy, when memory fills up, Valkey can remove that cache key to free up space, while retaining keys without a TTL.

One benefit: retain important data while making room when memory gets tight.

Learn more: bit.ly/4fpNzdh

Every application that talks to Valkey has to answer the same question in its test suite: where does the server come from? A real valkey-server in CI is accurate, but it adds a container, a port, and startup time to every run. A mock is fast, but it only knows what you told it, so it can drift from the server without failing a single test.

In a new post on the Valkey blog, Daniel Moran, walks through a third option: FakeValkey, an in-memory test double that runs the Valkey command set inside your Python test process. Learn more about FakeValkey and how to use it: bit.ly/4BPgIs9

Every conference has an afterparty, and this year Valkey is throwing a family reunion!

Join us for House of Valkey in Las Vegas on Tuesday, December 1, from 6 to 10 p.m. at HaSalon in The Venetian during re:Invent week. Come meet the builders, maintainers and contributors in town, grab some House of Valkey swag and get a family portrait. You don't need a re:Invent pass.

Space is limited, so request your spot early. Hosted by the @Percona and @AWS: luma.com/qi5jgr7d

In the latest Valkey blog, Khawaja Shams of @momentohq walks through the noisy neighbor problem that shows up in multi-tenant systems and media workloads, and shows how reply copy avoidance in Valkey 9.1 can fix it: valkey.io/blog/large-obj…

Registration closes today for Tuesday’s Valkey and @OpenSearchProj #meetup at @Uber's Sunnyvale office.

Production stories, vector search, what’s new in both projects, and pizza 🍕 Nov. 10, 5 to 8 p.m. PT

> luma.com/spa3weit

#Valkey #OpenSource