Redis with failover.
Flat rate.
A managed Redis cache/KV store with real high availability — master, replicas and Sentinel quorum behind one stable endpoint — created by API or your AI agent. The flat-rate, EU-hosted alternative to AWS ElastiCache.
ElastiCache is Redis with an AWS bill on top
ElastiCache prices per node-hour, per replica, plus data transfer, and you still design the failover topology and the endpoint routing yourself. Run a self-hosted Redis instead and you own the Sentinel setup, the failover testing, the backups and the TLS. koigrid gives you managed HA Redis — master + replica + 3 Sentinels + a master-aware endpoint + backups — as one API call, at a flat price with bandwidth included.
How koigrid compares
| koigrid | AWS ElastiCache | Self-hosted Redis | |
|---|---|---|---|
| High availability (auto-failover) | Included (Sentinel quorum) | Multi-AZ (extra) | You build it |
| Stable endpoint follows the master | Built-in (master-aware HAProxy) | Configuration endpoint | You wire it |
| Snapshots to object storage | Built-in + restore | Included | You script it |
| TLS | Included | Optional | You configure it |
| Data transfer | Included | Metered | VM cost |
| Pricing | Flat rate | Node-hours + replicas + transfer | VM + your time |
| Create/manage by API or AI agent | Native | Complex | SSH |
From public plans, 2026. koigrid HA = master + replica(s) with anti-affinity + 3 Sentinels for quorum + master-aware routing.
Real HA, not just a cache
Automatic failover
Master + replica on separate hosts with a 3-Sentinel quorum. Kill the master → a replica is promoted and the endpoint follows it.
Master-aware endpoint
One stable address; an HAProxy tcp-check keeps writes pointed at the current master through failovers.
Snapshots + restore
Daily RDB snapshots to EU object storage, plus on-demand — restore into a fresh cluster in one call.
TLS + auth
rediss:// with a per-cluster certificate and a password by default — safe over the public internet.
Self-healing
A reconcile loop follows the master via Sentinel and re-provisions any node that goes down — redundancy restores itself.
Agent-native
Create and manage caches from Claude Code or any LLM via the API — full dashboard parity.