talkincode/esmqtt repository overview

⁠Esmqtt

Esmqtt is a tool that can subscribe to receive messages through the MQTT protocol and forward them to Elasticsearch.

⁠Features

• Support for MQTT Protocol V3, V4
• Forward MQTT messages to Elasticsearch.
• Support message rule routing, customize message subject to correspond to the specified Elasticsearch indexes.
• Support message persistence

⁠Quick start

⁠Install the esmqtt service

go install github.com/talkincode/esmqtt

esmqtt -install

Use systemd to manage services on a Linux system

systemctl <start | stop | restart> esmqtt

⁠docker Deploy the ESMQTT service

docker-compose.yml⁠

```bash

docker-compose up -d

⁠Configuration

• yaml

appid: esmqtt
location: Asia/Shanghai
workdir: /var/esmqtt
debug: true
logger:
  mode: development
  console_enable: true
  loki_enable: false
  file_enable: true
  filename: /var/esmqtt/esmqtt.log
  queue_size: 4096
  loki_api: http://127.0.0.1:3100
  loki_user: esmqtt
  loki_pwd: esmqtt
  loki_job: esmqtt
  metrics_storage: /var/esmqtt/data/metrics
  metrics_history: 168
mqtt:
  server: ""
  username: ""
  password: ""
  debug: false
elastic:
  server: http://127.0.0.1:9200
  api_key: ""
  username: elastic
  password: elastic
  debug: false
  

• ENVIRONMENT VARIABLE

ESMQTT_SYSTEM_WORKER_DIR=/var/esmqtt
ESMQTT_SYSTEM_DEBUG=true

ESMQTT_MQTT_SERVER=tcp://127.0.0.1:1883
ESMQTT_MQTT_USERNAME=esmqtt
ESMQTT_MQTT_PASSWORD=
ESMQTT_MQTT_DEBUG=true

ESMQTT_ELASTIC_SERVER=https://localhost:9200
ESMQTT_ELASTIC_APIKEY=
ESMQTT_ELASTIC_USERNAME=elastic
ESMQTT_ELASTIC_PASSWORD=elastic
ESMQTT_ELASTIC_DEBUG=true

ESMQTT_LOGGER_JOB=esmqtt
ESMQTT_LOGGER_SERVER=
ESMQTT_LOGGER_USERNAME=esmqtt
ESMQTT_LOGGER_PASSWORD=
ESMQTT_LOGGER_MODE=development
ESMQTT_LOGGER_LOKI_ENABLE=false
ESMQTT_LOGGER_FILE_ENABLE=true

⁠Message rules

[
  {
    "topic": "testnode/elastic/message/create",
    "index": "testnode_message",
    "spliter": "day"
  }
  
]

• The program will load the file at startup, and if it does not exist, the program will automatically subscribe to the topic elastic/message/create. The index depends on the index field specified in the message.
• If neither the index field value is specified in the message nor in the rule, the message will be ignored.
• When using a rule, if no index field value is specified in the message, the index field value in the rule will be used. The value of spliter is "year | month | day | hour" (or null), and the index name is generated according to the date rule. For example testnode_message_2021-01-01, testnode_message_2021-01, testnode_message_2021.

⁠Message model

{
  "data": {
     "id": "1312313142",
     "index": "testindex",
     "payload": {
       "name": "test", "value": 100
     },
     "vector": [],
     "timestamp": 1698827090749
  }
}

• data.id specifies the document ID that the message is forwarded to, if not specified in the message, the internally generated UUID will be used.
• data.index specifies the index to which the message will be forwarded, if not specified in the message, the value of the index field in the rule will be used, if not specified in the rule, the message will be ignored.
• data.payload is a custom object, if it is empty, the message will be ignored.
• data.timestamp is the message timestamp, if it is empty, the current timestamp will be used.
• data.vector is the message vector, if it is empty, an empty array will be used (this field is designed for GPT model, with dimension 1536, for specific scenarios, it can be empty).