provider-cloudnative-pg is now an official OpenEverest provider
The community-built CloudNativePG provider — started as a personal project by Aditya Pimpalkar in June — has moved into the OpenEverest organization.
Latest news and articles from the OpenEverest project
The community-built CloudNativePG provider — started as a personal project by Aditya Pimpalkar in June — has moved into the OpenEverest organization.
The banner at the top of openeverest.io keeps inviting you to talk to a maintainer. This post explains who those calls are for, what happens on them, and how to book one.
We ran TiDB on an 8-node Kubernetes cluster with OpenEverest and TiDB Operator v2 and measured everything a self-hoster asks about: time to a working cluster, HTAP with TiFlash under CH-benCHmark, scaling under load, backups to S3, restores, upgrades and TLS.
Models served by OpenEverest no longer have to share one API key. With the User Management plugin, every application, team or person gets its own key for a model, with its own token quota, its own usage metrics, and rotation or revocation that affects nobody else.
OpenJev is an open-weights decision model: you ask a typed question, it answers with a probability for every option in one forward pass. I ran it on a $0.67/hour GPU with OpenEverest and KServe, let it triage 299 real KServe issues and diagnose 19 broken pods, and published it on the Envoy AI Gateway with API keys and token quotas. Here is what it got right, what it got confidently wrong, and the prompt fix that made things worse.
Meet Shivansh Sinha — a DevOps and backend engineer in Bengaluru who started by filing bugs against the OpenEverest v2 dev builds, and kept going until he was fixing them himself.
Milvus is now available on OpenEverest. I deployed it as a platform service, indexed the whole Kubernetes documentation for semantic search, upgraded it under live traffic, and cloned it into staging from a backup. Here is what that looked like.
I put a sharded Valkey cluster under full load, doubled the number of shards, shrank it back, drained a node and killed primaries. Here is what clients saw, how much data was lost, and what still needs fixing.
We added MaxScale to the MariaDB provider. I killed primaries in a few different ways to see how failover behaves, what clients see, and what still needs fixing.
Learn how Locci Cloud experimented with and eventually adopted OpenEverest to their platform.