Best practices for Tag Data Interface

QoS level

  • QoS 0: Highly recommended, since QoS 1 and 2 options can cause higher CPU utilization.

  • QoS 1 & 2: Not recommended, since data retention is default enabled for Tag Data Interface.

Tag Data Retention

This feature enables reader to buffer the tag events and stream data back to server in case of network issues or server failures.

  • Current retention configuration can be queried using GET /config API for Local REST deployment or get_config command for MQTT deployments. Retention is default enabled for Tag Data Interface.

  • Sample retention configuration:

    {
    "maxEventRetentionTimeInMin": 30,
    "maxNumEvents": 150000,
    "throttle": 100
}

    • throttle: Rate to which reader should send retained events to server. (default 500 TPS)

      • For large number of reader deployments it is recommended to adjust the throttle depends on the number of readers deployed, tag read rate and server processing capability.

      • throttle value should be greater that the read rate to reduce the deploy of current tag data.

    • maxNumEvents: Maximum number of events to be retained. (default 150000 Events)

    • maxEventRetentionTimeInMin: Single event retention time. (default: Events will be retained until reboot or maxNumEvents exceeded).

  • Retention configuration can be updated using PUT /config API for Local REST deployment or set_config command for MQTT deployments.

  • Retention can be disabled by configuring maxNumEvents to 0.

Tag Data Batching

This feature enables reader to group multiple tag data events into single event based on the configuration. Batching reduces network usage as well as reader CPU usage.

  • Current batching configuration can be queried using GET /config API for Local REST deployment or get_config command for MQTT deployments. Batching is default enabled for HTTP POST Endpoint and disabled for MQTT Endpoint.

  • Sample batching configuration:

    {
    "maxPayloadSizePerReport": 256000,
    "reportingInterval": 2000
}

    • reportingInterval: Batched event report interval in milli seconds. (default: 1000 milli seconds for HTTP POST Endpoint).

    • maxPayloadSizePerReport: Maximum batching event size in bytes. (default: 256000 bytes for HTTP POST Endpoint).

  • It is recommended to use batching for higher tag read rate scenarios to reduce network usage and reader CPU usage.

  • Batching configuration can be updated using PUT /config API for Local REST deployment or set_config command for MQTT deployments.

  • Batching can be disabled by configuring maxPayloadSizePerReport and reportingInterval to 0.