HiveMQ's Serverless retirement affects Pay As You Go clusters exactly like free ones: on December 31, 2026 they stop accepting connections and are later deleted. For teams that run real products on Pay As You Go, this is the moment to choose a home designed for production rather than another stopgap. The ConnectMQ Managed Cloud MQTT Broker gives you an MQTT service that ConnectMQ operates for you, with isolation and capacity that match a production fleet.
For the retirement details, read HiveMQ Cloud Serverless end of life.
Why Production Workloads Need More Than Serverless
Shared serverless clusters are excellent for getting started, and ConnectMQ's own Serverless plans support production fleets up to 1,000 connections. But many Pay As You Go users have requirements that point to a dedicated deployment:
- Predictable performance. On a dedicated cluster, no other tenant's traffic shares your broker, so latency depends only on your workload.
- Capacity planning. The cluster is sized for your fleet and growth plan, instead of a quota designed for the average tenant.
- Security reviews. "Does any other customer's traffic touch our broker?" is easy to answer when the answer is no.
- Fan-out and heavy consumers. Many backend services on shared subscriptions, larger payloads and QoS 2 flows are easier to accommodate.
The dedicated MQTT broker guide lists the signals that you have outgrown a shared tier.
Managed Cloud Options at ConnectMQ
| Option | Who runs it | Where it runs | Typical fit |
|---|---|---|---|
| Dedicated | ConnectMQ | An isolated cluster in ConnectMQ's cloud, only for you | Production fleets that want isolation without operating infrastructure |
| BYOC (Bring Your Own Cloud) | ConnectMQ | Inside your own cloud account | Data residency, existing cloud commitments, strict security boundaries |
| Self Managed Enterprise | You, with a licensed ConnectMQ broker | Your own servers or factory network | On-premises, air-gapped or regulated sites |
All three use the same MQTT 3.1.1 and 5.0 protocol support and the same portal concepts, so you can start on Dedicated and later add a BYOC or on-premises site without redesigning topics or clients. Read more about BYOC MQTT brokers and the enterprise MQTT broker options.
What Stays the Same After You Move
For a Pay As You Go fleet, the protocol-level experience is deliberately unchanged:
- Standard MQTT clients and libraries, with no proprietary SDK.
- TLS on port 8883 and secure WebSocket on 8084 (path
/mqtt), with publicly trusted certificates. - Your topic structure, payload formats and QoS levels.
- MQTT 5 features such as reason codes, user properties, message expiry and shared subscriptions.
What changes is the hostname, the credentials and the client ID scheme, plus who you call when something looks wrong.
Tooling for Operations Teams
A production broker is only as good as your visibility into it. The ConnectMQ portal gives operations teams:
- A dashboard of connections and traffic.
- A live topic explorer and an in-browser MQTT client for debugging.
- An ISA-95 device view with Sparkplug B decoding, and optional republishing of Sparkplug B data into ISA-95 topics for a Unified Namespace.
- Threshold alert rules and an alerts log.
- Connection state per client on
$SYS/broker/connection/{clientId}/state, which your own monitoring can subscribe to.
For industrial users this matters: see MQTT Sparkplug B and Unified Namespace for how ConnectMQ handles ISA-95 topic structures.
Plan the Move as a Production Change
A production migration deserves a change plan rather than a weekend scramble:
- Capacity. Measure your peak concurrent connections and message rate on HiveMQ and add headroom for 12 to 18 months of growth.
- Environment. Ask for the Dedicated or BYOC deployment early so it is provisioned well before your first wave.
- Staging. Point a staging fleet or a test rig at the new broker and run your load profile. The guide on MQTT load testing tools helps.
- Dual-run. Relay traffic between HiveMQ and ConnectMQ while you move consumers, then devices, in waves. See zero-downtime HiveMQ migration.
- Cutover and review. Move the last devices well before December 31, 2026, then review metrics and alerts for a few weeks before deleting the old cluster.
Start Today, Upgrade When Ready
You do not have to wait for a dedicated cluster to begin. Create a free broker at app.connectmq.com, validate your client libraries, TLS settings and topic structure, and contact ConnectMQ through connectmq.com to size a Dedicated or BYOC deployment. Then follow the HiveMQ Serverless migration guide for the move itself.
Frequently Asked Questions
What is a Managed Cloud MQTT Broker?
It is an MQTT broker that the provider operates for you on infrastructure reserved for your organisation. ConnectMQ handles servers, upgrades, certificates and monitoring of the broker, while you keep full control of your clients and data flows.
Should HiveMQ Pay As You Go users choose Serverless (Paid) or Dedicated?
If your fleet stays below 1,000 concurrent connections and you are comfortable with a shared, isolated cluster, Serverless (Paid) is the simplest move. Choose Dedicated when you need isolation, predictable latency, more capacity or a cleaner security review.
Can ConnectMQ run inside my own cloud account?
Yes. With BYOC, ConnectMQ manages the broker inside your cloud account, which helps with data residency and existing cloud commitments.
How long does it take to move a production fleet?
Plan for several weeks: provisioning, staging tests, dual-run and staged waves. The deadline is December 31, 2026, so start as early as possible.
