// Managed Kafka
Managed Kafka on the clusters you already run.
A Kafka managed service for your own infrastructure, in any cloud or none.
OSSeva Operate is a managed Kafka service for self-managed Apache Kafka: our engineers run your brokers 24/7 on your own servers, cloud account or Kubernetes, and OSSeva does not host them. Amazon MSK, Confluent Cloud and Google Cloud's Managed Service for Apache Kafka run Kafka inside the provider's service, and Azure Event Hubs offers a Kafka-protocol endpoint on its own platform. OSSeva covers patched Kafka 2.8.x to 3.9.x, current 4.x in KRaft mode, and Kafka Connect and Streams, priced per cluster.
Trusted globally by enterprises
Why teams want Kafka managed on their own infrastructure
Cloud Kafka services are a good answer for many teams. These are the reasons others keep their clusters and hand over the operations instead.
Your version is older than the managed service allows
Amazon MSK's own table lists end of support for Kafka 3.6.0 on 1 June 2026 and 3.7.x on 1 September 2026, and Kafka 4.2 on MSK runs on Express brokers only. A self-managed cluster on 2.8 to 3.7 has to upgrade before it can move, and a ZooKeeper cluster also has to reach KRaft before 4.x.
A Kafka endpoint is not the whole of Kafka
Azure Event Hubs speaks the Kafka protocol without Kafka brokers. Microsoft's documentation lists Kafka Streams and transactions as public preview on the Premium and Dedicated tiers and gzip as the only compression, and points to native Kafka on Azure HDInsight for features the endpoint lacks.
The data and the network stay where they are
Kafka usually sits in the middle of everything: databases, services, connectors and downstream analytics. Moving it into a provider's service changes networking, identity and residency for every producer and consumer at once.
Connect, Streams and consumer lag need operating too
Most Kafka incidents surface as consumer lag, a failing connector or an unbalanced partition leader, not a dead broker. Someone has to watch for them at any hour and know the applications behind them.
The dates that matter
2025-03-18
Apache Kafka 4.0.0 released. It runs only in KRaft mode, so ZooKeeper clusters migrate on 3.x first.
2025-04-21
Amazon MSK adds Kafka 3.9, the last line with both ZooKeeper and KRaft, with MSK support for at least two years from that release.
2026-06-01
Amazon MSK end of support for Kafka 3.6.0.
2026-07-15
Amazon MSK adds Kafka 4.2, on Express brokers only.
2026-09-01
Amazon MSK end of support for Kafka 3.7.x.
What OSSeva delivers
24/7 broker and consumer-lag monitoring
Consumer lag, partition leader balance and broker JVM metrics watched around the clock, with runbooks written for your clusters before we take the pager. P1 incidents get a 15-minute response, and capacity planning and scaling are part of the service.
Patched Kafka, older lines included
Signed builds for Kafka 2.8.x to 3.9.x after upstream stops fixing them, and current 4.x, in ZooKeeper or KRaft mode. OSSeva Operate patches Critical CVEs (CVSS ≥ 9.0) within 48 hours and High within 7 days.
Connect, Streams and the move to KRaft
Operational support for Kafka Connect and Kafka Streams, with connector compatibility checked against patched brokers. Rolling upgrades run broker by broker, and the ZooKeeper to KRaft migration is planned and run with your team.
Your options, compared
| Option | What you get | Trade-off |
|---|---|---|
| Amazon MSK | AWS runs the control plane and replaces failed brokers. MSK Provisioned (Standard or Express brokers) or MSK Serverless, on open-source Apache Kafka. | AWS only. Versions follow MSK's support table: 3.6.0 and 3.7.x reached MSK end of support in 2026. |
| Confluent Cloud | A fully managed data streaming platform on AWS, Google Cloud and Azure, with Confluent's own Kafka engine, stream processing and governance. | Runs in Confluent's service, and leaving later means a cluster migration. |
| Azure Event Hubs (Kafka endpoint) | A Kafka-protocol endpoint on a fully managed Azure service, on the Standard, Premium and Dedicated tiers. Existing clients often need only a configuration change. | Not Kafka brokers. Streams and transactions are public preview on Premium and Dedicated, and compression is gzip only. |
| Google Cloud Managed Service for Apache Kafka | Google runs open source Apache Kafka clusters and handles broker replacement, scaling and security patching. | Google Cloud only, and the clusters live in Google's service. |
| OSSeva Operate | 24/7 operations for self-managed Kafka on your servers, any cloud account or Kubernetes, with patched 2.8.x to 3.9.x and current 4.x. | You keep owning and paying for the infrastructure. Priced per cluster. |
Version dates from the Amazon MSK Developer Guide and the Apache Kafka blog. Provider descriptions from the Confluent Cloud, Azure Event Hubs and Google Cloud documentation. Checked 9 October 2026.
Frequently asked questions
Do you host Kafka for us?
No. OSSeva does not host Kafka. We operate the clusters you run on your own servers, VMs, cloud account or Kubernetes, through the access you grant, and the brokers and data stay in your account. If you want no infrastructure at all, a cloud Kafka service is the simpler choice.
How does managed Kafka from OSSeva compare with Amazon MSK?
MSK runs open-source Apache Kafka inside AWS's service, with AWS handling the control plane and broker replacement, on the versions in MSK's support table. OSSeva Operate runs the self-managed Kafka you already have, in AWS or anywhere else, including versions MSK no longer supports, and our engineers handle monitoring, incidents, patching and upgrades.
Is Azure Event Hubs a managed Kafka service?
Event Hubs provides a Kafka endpoint on a fully managed Azure service, on the Standard, Premium and Dedicated tiers. It implements the Kafka protocol rather than running Kafka brokers, and Microsoft points to native Kafka on Azure HDInsight for features the endpoint lacks. OSSeva can also operate Apache Kafka on VMs or Kubernetes inside your own Azure subscription.
Is there a managed Kafka service on Google Cloud?
Yes. Google Cloud's Managed Service for Apache Kafka runs open source Apache Kafka clusters, with Google handling broker replacement, scaling and security patching. If you run Kafka on Compute Engine or GKE yourself, OSSeva Operate can take on the operations there instead.
When is a cloud provider's managed Kafka the better fit?
When you are in one cloud, your version and features fit what the service offers, and you would rather not own brokers at all. OSSeva Operate suits teams that run Kafka across several environments, keep clusters on premises, or run versions a managed service no longer supports.
Which Kafka versions can OSSeva Operate run?
Kafka 2.8.x to 3.9.x with patched builds after upstream stops fixing them, and the current 4.x line in KRaft mode. Both ZooKeeper-mode and KRaft-mode clusters are covered, along with Kafka Connect and Kafka Streams.
How fast are Kafka CVEs patched under managed operations?
OSSeva Operate patches Critical CVEs (CVSS ≥ 9.0) within 48 hours and High within 7 days. Fixes arrive as signed builds of the Kafka version you already run and are rolled out broker by broker in your maintenance windows.
How is managed Kafka from OSSeva priced?
OSSeva Operate for Kafka is priced per cluster, not per broker, partition or throughput tier. Kafka and the rest of your open source stack sit on one contract with one renewal date. Book a discovery call for a quote.
Keep your Kafka. Hand us the pager.
Discovery call, cluster inventory, then a quote priced per cluster.