Your transactions, as a stream.

You could, and plenty of people do. Then the patching, the upgrades, the monitoring and the restores are yours as well. We run the broker for you, on its own or wired into your systems by our data engineers, next to what you run with us. Out-of-hours cover comes with the paid advanced support plan.

We run Apache Kafka itself, with KRaft, not an API-compatible substitute. The same goes for NATS, RabbitMQ, ActiveMQ, and Eclipse Mosquitto for MQTT.

Yes. Each is provisioned on instances of its own, so a Kafka cluster for application events and an MQTT broker for device telemetry sit side by side without either knowing the other exists.

Yes. We peer directly with AWS, Azure, GCP, Oracle and Cloudflare at European exchange points, so a producer or consumer in one of them reaches the broker without crossing the open internet. Our engineers can also point Debezium at a database you run there or on your own premises, as long as it reaches us over direct peering or Private Link and the connection uses TLS. We do not read a database that is open to the public internet. You do not have to move anything first. Latency from a public-cloud region is whatever that distance costs; single-digit milliseconds is a same-site figure.

Data-Driven is for analytics on a schedule: pipelines, a lake and the reports built on it. This is for events the moment they happen, moving between the systems you run. The two compose: our engineers can land your events in the lake Data-Driven builds. If what you need is a report every morning, start with Data-Driven, and the discovery call will say so.

Patching and version upgrades, monitoring and alerting, backup and restore, and on a cluster, failover within the datacenter you chose, at 99.9% uptime, by contract, with automatic credits if we miss it. Support is free on every product we sell, with a 4-hour target response time on support requests, and the paid advanced plan adds shorter targets plus weekend and out-of-hours cover. A broker replicated across more than one site is not included: we design and build that with you, over our own Layer 2 network.

Single-digit milliseconds when your producers and consumers run in the same datacenter as the broker. From an office or a plant over Private Link, it is whatever that distance costs. Replication to another site crosses our own Layer 2 network rather than the public internet, but it is still a real distance, so we do not publish a number for it: a single figure could fit one pair of sites and would mislead about the rest.

Yes, and it is not a line item. Backups run hourly and are kept on a grandfather-father-son schedule: the last 24 hourly, 7 daily, 4 weekly, 12 monthly and 3 yearly. We do not publish a restore time yet.

The discovery call, and it is free. 60 minutes, an engineer rather than a salesperson, and a written diagnosis inside a week: which broker, what it connects to, and what we would leave alone. Already a customer? Same call, same form: tell us what you run with us and it starts from there.

No. Start with one topic, usually the thing already causing pain, like an order or a transaction stream, and add brokers and topics as you need them.

Madrid, Munich, Helsinki or Amsterdam, and you pick which at the start. It is our own hardware in our own datacenters. The broker runs in that site, and in a second one only if you ask us to build a replica there. Replication between our sites runs over our own Layer 2 network and never leaves the EU.

DIVERSITY is a European company with no parent company, US or otherwise, so there is no entity above us for an order from outside the EU to be served on. That is a structural answer rather than a policy one.