Sergey is the founder of Solanica and a product leader focused on delivering robust open-source solutions. After consistently contributing to the growth of Percona and its open-source product culture as a VP of Product, he is now one of the maintainers of OpenEverest, the CNCF open-source data and AI platform for Kubernetes. A digital nomad who enjoys spreading the word about data on Kubernetes. He loves playing tennis and hockey, but his major passions are his family, three kids and a dog.
We broke PostgreSQL 18 and MySQL 8.4 on Kubernetes 22 different ways - OOM kills, failed probes, bad Secrets, PVC and disk problems, config typos and version changes - and wrote down exactly what CrashLoopBackOff looked like each time, how to diagnose it in a minute, and which fixes actually work. Several 'fixed' pods came back Ready but broken.
We ran 9 memory-failure scenarios against real Valkey Instances on a live managed Kubernetes cluster: no maxmemory, maxmemory at the limit, BGSAVE, AOF, fragmentation, QoS classes, and a replica full resync that quietly erased every key while the cluster kept reporting itself healthy.
Sovereign S3 is more than storing data in the right country. Learn the four pillars of object storage sovereignty and build a self-hosted, MinIO-free S3 on Kubernetes with SeaweedFS and OpenEverest.
We put a Valkey semantic cache in front of a RAG pipeline on Kubernetes - Qwen on vLLM, MiniLM embeddings, Milvus, all deployed as OpenEverest Instances - and measured it. Cache hits answered in 88 ms instead of 4.7 s and cut LLM tokens by three quarters. Most of that came from exact repeats, and every bit of extra fuzziness bought wrong answers.
We ran pod deletes, node drains, hard kills and network partitions against a 3-replica SQL Server Availability Group on Kubernetes while a probe wrote every 100ms. Here is what the testing surfaced, what we changed, and the result: 20 out of 20 fault runs with zero acknowledged writes lost.
Run your own OpenAI-compatible endpoint on your GPUs with OpenEverest v2.0.0-dev.3 and the KServe provider - per-model API keys, token quotas, HTTPS, and models that span several GPU nodes.
OpenEverest v2.0.0-dev.3 introduces Presets - reusable templates for instances and clusters that turn database deployment into a single click or a simple API call.
Solanica's MSSQL operator and OpenEverest provider now ship SQL Server 2025 as the default engine and deliver provider-managed point-in-time recovery (PITR).
How Solanica Platform runs SQL Server on Kubernetes: a Go operator with Always-On AGs, Lease-based failover, native object-storage backups, and an OpenEverest provider.
