Private Cloud subscription
TBMQ Private Cloud provides dedicated instances based on a flexible, pay-as-you-go model. The interactive pricing calculator lets you define your exact capacity so you only pay for the performance and scale you need. Once your core capacity is set, your private broker instance operates under the following default and configurable limits. The values are monthly limits unless stated otherwise.
| Parameter | Value | Description |
|---|---|---|
| Sessions | From 5K to Unlimited | Maximum number of client sessions managed by the broker, including both active connections and persistent sessions that are temporarily disconnected but still maintained for quick reconnection. Once a session expires or is explicitly removed, it no longer counts toward the limit and the slot becomes available for a new session. |
| Throughput (msg/sec) | From 1K to Unlimited | Total throughput of messages processed per second, including both incoming and outgoing MQTT_PUBLISH packets. Does not include acknowledgment packets used in QoS 1 and 2 flows (PUBACK, PUBREC, etc.), nor other MQTT control packets (CONNECT, SUBSCRIBE, UNSUBSCRIBE, etc.). |
| Monthly traffic | From 200 GB to Unlimited | Total data transmitted through the broker, including both incoming and outgoing MQTT messages. Traffic is measured as the sum of all MQTT packet payloads sent and received by clients. If the monthly allowance is exceeded, additional usage is billed at $0.1 per GB. |
| Application topic size | 1 GB | Maximum size of the Kafka topic used to store persisted messages for a disconnected APPLICATION client. Each client has a dedicated topic; messages are retained until delivered or the size limit is reached. |
| Application topic TTL | 1 day | Time-to-live for messages stored in Kafka topics for APPLICATION clients. If a client remains disconnected beyond this period, undelivered messages are discarded automatically. |
| Integration topic size | 1 GB | Maximum size of the Kafka topic used to store persisted messages for a disconnected Integration. Each Integration has a dedicated topic; messages are retained until delivered or the size limit is reached. |
| Integration topic TTL | 1 day | Time-to-live for messages stored in dedicated Kafka topics for each Integration. If the Integration remains disconnected beyond this period, undelivered messages are discarded automatically. |
| Messages storage for DEVICE persistent sessions | 100 per session | Maximum number of messages stored for each DEVICE client with a persistent session while it is offline. Once the limit is reached, older messages are dropped to make room for new ones. |
| Device persisted messages TTL | 1 day | Time-to-live for messages stored for DEVICE persistent clients while they are offline. |
| Stats TTL | 1 month | Time-to-live for statistics persistence. Includes metrics visible in the UI: sessions, subscriptions, incoming messages, outgoing messages, etc. |
| Session TTL | 1 day | Time-to-live for a disconnected persistent session. If a client remains disconnected beyond this duration, its session data is automatically removed from the broker. |
| Client incoming messages rate limit | Up to 50/sec, not exceeding 3,000/min | Total number of incoming messages per publisher client with any QoS level. |
| Client outgoing messages rate limit | Up to 50/sec, not exceeding 3,000/min | Total number of outgoing messages per non-persistent subscriber client with QoS 0 (AT_MOST_ONCE). |
| In-flight messages per publisher | 100 | Maximum number of incoming QoS 1 and QoS 2 messages from a publisher awaiting acknowledgment. Acknowledgment is sent after the message is successfully persisted by the broker. If this limit is exceeded, the client is disconnected to prevent overload. |
| Client pre-connect messages | 100 | Maximum number of messages the broker queues per client during the connection handshake phase. If this limit is exceeded before the client completes the connection, the client is disconnected. |
| Message size | 1 MB | Maximum payload size in an MQTT_PUBLISH packet that the broker will accept. Larger messages are rejected. |
| Uptime SLA | 99.9% or 99.99% | Percentage of time the broker is expected to be fully operational and accessible. |
Where K = 1 thousand, M = 1 million, B = 1 billion.