On this page
Short answer
If the workload is standard, already in a public cloud and not bound by data residency, a hosted service is the fastest route: Confluent Cloud for the fullest Kafka feature set and connectors, Amazon MSK for AWS-native billing and IAM. If the data cannot leave your account, the estate is customised, or you run components the platform does not offer, self-managed Kafka with a support contract keeps control and adds an engineer when it breaks. Operated for you on your infrastructure sits between: your account, your topology, someone else's pager. Running it yourself with no contract is right only for teams that have done it before and have the on-call rota to prove it. AceMQ provides the operated-for-you option as managed Kafka.
Four ways to run Kafka, on the same criteria
| Confluent Cloud | Amazon MSK | Self-managed with a support contract | Operated for you (AceMQ managed operations) | |
|---|---|---|---|---|
| Where the cluster lives | Confluent's account, in your chosen cloud and region | Your AWS account, as a managed service | Wherever you put it | Your infrastructure or cloud account |
| Who chooses the version | Confluent | AWS, from its supported list | You | You, with an upgrade path planned and executed for you |
| Connectors and components | Confluent's catalogue and Schema Registry, ksqlDB, Stream Governance | MSK Connect and the open-source ecosystem; check each connector | Anything Apache Kafka supports | Anything Apache Kafka supports |
| Who answers an incident | Confluent support tiers | AWS support tiers for the service; your team for topics, consumers and connectors | Your on-call, with the contract as escalation to a senior engineer | Named senior engineer, 24/7, 15-minute emergency SLA |
| Cost shape | Usage-based; throughput, storage, connectors and partitions billed separately | Broker-hours and storage on the AWS invoice | Infrastructure plus a contract priced on cluster count and response tier | Engineer time; more than a support contract, without infrastructure leverage |
| Data residency and compliance | Confluent's account; check region and attestations | Your account and region | Unchanged | Unchanged |
| Exit | Mirror out; connectors and governance features do not travel | Mirror out; MSK-specific IAM auth does not travel | Nothing to exit | Runbooks and monitoring stay yours; exit is a handover |
| Wins when | You want the fullest managed feature set and are not residency-bound | All-AWS estate, standard workload, procurement simplicity | Custom estate, residency constraint, or a capable platform team | No Kafka specialist to spare and the data cannot move |
Confluent Cloud
The fullest managed Kafka: Confluent runs the cluster in its account in your chosen cloud and region, and sells the surrounding platform with it: Schema Registry, a connector catalogue, ksqlDB, stream governance, and tiered storage. Support follows the product and is authoritative on it. Cost is usage-based, which is easy to start and hard to forecast; throughput, storage, connectors and partitions are billed on separate lines. It wins when you want the platform, not just the broker, and are not bound by where the data lives.
Amazon MSK
Kafka as an AWS service inside your own account: brokers on a supported version list, IAM authentication, CloudWatch metrics, and billing on the AWS invoice. AWS supports the service; your topics, consumers, connectors and the incidents that come from them are yours, which is why MSK estates so often carry an independent support contract at the Kafka layer. It wins for an all-AWS estate with a standard workload and a procurement team that prefers one vendor.
Confluent vs Amazon MSK, head to head
Both are fully managed Apache Kafka in the cloud. The difference is how much of the data streaming platform around the brokers comes with the service, and who you buy it from.
| Confluent Cloud | Amazon MSK | |
|---|---|---|
| What you buy | A streaming platform, with Kafka as one part | Apache Kafka as an AWS service |
| Schema registry | Managed Confluent Schema Registry | AWS Glue Schema Registry, or Confluent Schema Registry run yourself |
| Connectors | Managed connectors in the same console | MSK Connect, running Kafka Connect connectors you choose |
| Stream processing | Managed, inside the platform | Whatever you deploy alongside |
| Clouds | AWS, Azure and Google Cloud | AWS only |
| Vendor | Confluent, an IBM company since 17 March 2026 | AWS, on the same invoice as the rest of the estate |
| Who owns consumer lag, partitions and client behaviour | You | You |
Choose MSK when the estate is all-AWS and the team is comfortable owning Kafka Connect and schema management. Choose Confluent Cloud when you want those managed as one product, or need Kafka across more than one cloud. The last row is why both sets of customers still carry Kafka-layer support: neither service diagnoses a rebalance storm in your consumers.
Self-managed Kafka with a support contract
You run the cluster; a senior engineer is on the other end of the contract when it breaks, and alongside you for upgrades, KRaft migration and capacity. Open-source Apache Kafka is fully supportable this way, and so is Confluent Platform. It keeps version choice, connector freedom and data residency entirely yours, at a cost well below a managed service. It fits a platform team that can hold the pager and wants expertise rather than outsourcing.
Operated for you, on your infrastructure
The cluster stays in your account, on your VMs or Kubernetes, and a specialist takes over monitoring, patching, upgrades, capacity and on-call. Nothing about data residency changes. It costs more than a support contract, because it is engineer time with no infrastructure leverage, and it exists for estates where nobody on the team wants to own Kafka and the data cannot move to a hosted platform.
Self-hosted Kafka vs cloud: what actually decides it
The self-hosted versus cloud question is usually framed as cost, and cost is the least reliable way to answer it. Four things decide it more cleanly. Where the data must live: a residency or air-gap obligation rules out a vendor-account service before price is discussed. Who holds the pager: self-hosting is cheap until the first multi-hour incident with nobody who has seen that failure before. How predictable the load is: usage-billed cloud Kafka is cheapest for spiky or small workloads and gets expensive on steady high throughput, which is where self-hosted hardware is at its best. How much of the platform you use: if you need Schema Registry, managed connectors and governance, you are buying a platform and cloud wins; if you need brokers, you are renting machines at a markup.
There is a middle that the question hides: self-hosted Kafka with someone else operating it, or with a support contract behind your own team. Both keep the data and the version choice in your account and remove the 3am problem.
How to decide in one pass
- Can the data leave your account? If not, Confluent Cloud is out; MSK stays in if you are on AWS.
- Do you need Confluent's platform components, or just Kafka? If the platform, Confluent Cloud or Confluent Platform. If just Kafka, any of the other three.
- Is there a platform team that can hold the pager? If yes, self-managed with a support contract. If no, a hosted service or operated-for-you.
- Is the estate customised, or on a version a platform will not run? If yes, self-managed or operated-for-you.
Frequently asked questions
What is managed Kafka, and which services offer it?
Managed Kafka, or Kafka as a service, means a provider runs the Kafka cluster for you: brokers, storage, upgrades across Kafka versions, scaling and recovery. The main fully managed Kafka services are Amazon Managed Streaming for Apache Kafka (Amazon MSK, with Standard and Express brokers or MSK Serverless), Google Cloud Managed Service for Apache Kafka, Confluent Cloud, Aiven for Apache Kafka, and the Kafka endpoint in Azure Event Hubs. Existing Kafka clients and Kafka topics carry over because these services run, or speak the protocol of, Apache Kafka. Running Apache Kafka yourself, self-managed on virtual machines or Kubernetes, keeps every setting and the cost of the infrastructure in your hands and makes the operations your team's job, or the job of a support or managed-service partner.
Is self-hosted Kafka better than cloud Kafka?
Neither is better in general. Self-hosted Kafka wins when data must stay in your own account, when throughput is high and steady, or when you need version and connector freedom. Managed cloud Kafka wins for small or spiky workloads, for teams with nobody to hold the pager, and when you want the surrounding platform such as Schema Registry and managed connectors. A third option is self-hosted Kafka that a specialist operates or supports for you.
Is Confluent Cloud or Amazon MSK better?
Confluent Cloud for the fullest managed feature set, connector catalogue and stream governance, in any major cloud. MSK for an all-AWS estate that wants Kafka on the AWS invoice with IAM. If the data cannot leave your account, MSK or self-managed; if you need Confluent's components, Confluent.
How does Amazon MSK pricing compare with Confluent Cloud pricing?
They meter different things. Amazon MSK provisioned clusters charge an hourly rate per broker instance, billed per second, plus storage per GB-month; AWS lists, for example, $0.204 an hour for a kafka.m7g.large Standard broker, and Express brokers add a per-GB charge for data written. MSK Serverless charges per cluster-hour, per partition-hour and per GB written and read, and MSK Connect charges per MSK Connect Unit hour. Confluent Cloud charges per eCKU-hour plus per GB in and out and per GB-month stored, with Standard clusters starting at about $385 a month and Enterprise at about $895 on Confluent's pricing page. Both vary by region; figures checked 7 October 2026. MSK usually looks cheaper per unit, while Confluent Cloud meters managed connectors, Schema Registry and Flink as separate extras, so compare the whole bill at your throughput rather than the cluster line.
Does AWS support cover Kafka problems on MSK?
AWS supports the MSK service: broker availability, the managed control plane. Your topics, consumer groups, connectors and the incidents they cause are outside that boundary, which is why MSK estates commonly add independent Kafka support.
Can I get support for self-managed open-source Kafka?
Yes. Apache Kafka is fully supportable as open source under an independent contract, AceMQ's included, with 24/7 cover and a 15-minute emergency SLA, and the same contract can cover MSK and Confluent Platform clusters.
What does it cost to have Kafka operated for you?
More than a support contract, because it is engineer time with no infrastructure leverage, and it is priced on cluster and broker count plus the response tier. The trade is a named engineer holding the pager on an estate that stays in your account.