HyperbyteDB Cloud
Point Telegraf at a URL. We run everything behind it.
Managed HyperbyteDB. The same InfluxDB v1 wire protocol, the same InfluxQL, the same core you can already pull from GitHub — with the tuning, replication, upgrades and capacity planning handled by the people who wrote it.
We're taking on a small number of design partners before general availability.
The numbers we'd put in your capacity plan
| Sustained ingest | ~1.5M points/sec at p95 under 250ms |
| Peak ingest | ~1.9M points/sec |
| Measured on | AMD EPYC 7413, 24c/48t, 251GB RAM, single node |
The engine on its own does 99 million rows a second. Put a network, a parser, a series index and a durable write-ahead log in front of it and you get 1.5 million. Both numbers are true; only one of them survives contact with production. We publish the second one.
Keep the stack. Change the URL.
- Same
/writeand/queryendpoints - Same InfluxQL — aggregates, derivatives, moving averages,
GROUP BY time(), subqueries, regex measurements - Same Telegraf, same Grafana datasource, same clients
- Continuous queries and HA replication included, not gated behind a sales call
Why teams hand this over
- Cardinality that doesn't page you. Wide
WHEREclauses and high-cardinality tags without the series explosion that kills InfluxDB v1. - Replication without a procurement cycle. InfluxDB charges enterprise pricing for master-master. It's in the box here.
- Upgrades, backups and restore tests run by the people who wrote the storage layer.
- More raw data on less hardware — so retention is a decision about what you need, not what you can afford to keep.
Self-hosted or managed
| Self-hosted | Cloud | |
| Licence | Open source, on GitHub | Early access |
| InfluxDB v1 compatibility | Full | Full |
| Replication & HA | Yours to run | Managed |
| Upgrades, backups, restore tests | Yours to run | Managed |
| Support | Community | Direct |
| Price | Free | Announced at general availability |
What we haven't built yet
Fair questions you'd rather hear from us than find out later. Regions, a published uptime SLA, formal compliance certification and bring-your-own-cloud are all still ahead of us, not behind us. Early access exists partly so design partners get a say in which of them lands first.
Request early access
Six questions, about two minutes. We'll come back with what we can actually commit to for your workload.