Message queue

ThingsBoard Edge uses a message queue for two purposes: internal communication between services (Transport → Rule Engine → Core) and the cloud event queue used for synchronization with the ThingsBoard server.

Queue backends

Edge supports two queue implementations. Set the backend with the TB_QUEUE_TYPE environment variable:

Backend	`TB_QUEUE_TYPE`	Persistence	Use case
In-Memory	`in-memory`	None — messages lost on restart	Development, single-node demos
Apache Kafka	`kafka`	Disk-based, survives restarts	Production, cluster deployments

Cloud event queue

The cloud event queue is separate from the inter-service message queue. It stores telemetry, attribute updates, alarms, and other events destined for the ThingsBoard server before they are pushed over gRPC.

Queue type	Storage location	Available since
PostgreSQL	`cloud_events` table	All versions
Kafka	`ts_cloud_event` topic	Edge 3.9

When TB_QUEUE_TYPE=kafka, the cloud event queue automatically uses the Kafka topic instead of the PostgreSQL table. This improves throughput and provides better durability for high-volume synchronization scenarios.

  ┌─────────────────────────────────────────────────────┐
  │  ThingsBoard Edge                                   │
  │                                                     │
  │  Rule Engine → push to cloud node                  │
  │       │                                             │
  │       ▼                                             │
  │  ┌──────────────────────────────────────────┐       │
  │  │  Cloud Event Queue                       │       │
  │  │  PostgreSQL (cloud_events table)         │       │
  │  │  — OR —                                  │       │
  │  │  Kafka (ts_cloud_event topic)  [3.9+]    │       │
  │  └─────────────────────┬────────────────────┘       │
  │                        │                            │
  └────────────────────────┼────────────────────────────┘
                           │  gRPC (async)
                           ▼
  ┌─────────────────────────────────────────────────────┐
  │  ThingsBoard Server                                 │
  └─────────────────────────────────────────────────────┘

The Cloud Manager reads events from the queue and pushes them to the server in the order they were stored. If the connection is lost, events continue to accumulate and are delivered once the connection is restored.

Cloud event delivery tuning

Variable	Default	Description
`CLOUD_RPC_STORAGE_MAX_READ_RECORDS_COUNT`	`50`	Records per uplink batch
`CLOUD_RPC_SLEEP_BETWEEN_BATCHES`	`1000` ms	Pause between consecutive batches
`CLOUD_RPC_NO_READ_RECORDS_SLEEP`	`1000` ms	Pause when the queue is empty
`CLOUD_RPC_MISORDERING_COMPENSATION_MILLIS`	`60000` ms	Extra wait for out-of-order events before sending

Message queue vs Rule Engine queues

The message queue described on this page is the infrastructure-level transport — Kafka or in-memory — that carries messages between Edge services (Transport, Rule Engine, Core).

This is different from Rule Engine Queues, which are a higher-level abstraction configured in the ThingsBoard UI. Rule Engine Queues control submit strategy (burst, batch, sequential), retry policy, polling interval, and parallelism. They run on top of the message queue backend.

Kafka configuration

When running with TB_QUEUE_TYPE=kafka, the following environment variables configure the connection and behavior.

Connection

Variable	Default	Description
`TB_KAFKA_SERVERS`	`localhost:9092`	Kafka broker address(es), comma-separated

Producer tuning

Variable	Default	Description
`TB_KAFKA_ACKS`	`all`	Acknowledgment level. `all` = wait for all in-sync replicas (safest). `1` = leader only (faster).
`TB_KAFKA_BATCH_SIZE`	`16384`	Maximum batch size in bytes before sending to broker
`TB_KAFKA_LINGER_MS`	`1`	Milliseconds to wait for more records before sending a batch. Higher values increase throughput at the cost of latency.
`TB_KAFKA_COMPRESSION_TYPE`	`none`	Compression algorithm: `none`, `gzip`. Use `gzip` to reduce network bandwidth.
`TB_KAFKA_RETRIES`	`1`	Retry count for transient send failures
`TB_KAFKA_MAX_REQUEST_SIZE`	`1048576`	Maximum request size in bytes (1 MB default)

Consumer tuning

Variable	Default	Description
`TB_QUEUE_KAFKA_MAX_POLL_RECORDS`	`8192`	Maximum records returned per `poll()` call
`TB_QUEUE_KAFKA_MAX_POLL_INTERVAL_MS`	`300000`	Maximum time between `poll()` calls before the consumer is considered dead (5 minutes)
`TB_QUEUE_KAFKA_SESSION_TIMEOUT_MS`	`10000`	Heartbeat session timeout
`TB_QUEUE_KAFKA_AUTO_OFFSET_RESET`	`earliest`	Where to start reading when no committed offset exists: `earliest` or `latest`

In-memory queue

The in-memory queue stores all messages in JVM heap memory. No configuration is needed beyond TB_QUEUE_TYPE=in-memory. Messages are lost if the Edge process restarts — including cloud events that have not yet been pushed to the server.