LOG Indexing Service is a Druid plug-in launched by E-MapReduce and is used to consume data from Log Service.


LOG Indexing Service consumes data in a similar way as Kafka Indexing Service and supports exactly-once semantics. LOG Indexing Service has advantages of both Log Service and Kafka Indexing Service.
  • Provides various convenient methods for collecting data to Log Service.
  • Eliminates the need for Kafka clusters, which shortens the path of the data flow.
  • Supports exactly-once semantics.
  • Retries failed jobs to ensure reliability for data consumption, and allows for cluster restarts and service updates without service interruption.


  • Make sure that you have activated LOG and have configured projects and logstores.
  • Prepare the following configuration items:
    • The endpoint of LOG. Use the intranet endpoint.
    • A pair of AccessKey ID and AccessKey Secret to access LOG.

Use LOG Indexing Service

  1. Prepare the ingestion spec.
    LOG Indexing Service is similar to Kafka Indexing Service. For more information, see Kafka Indexing Service. The same data is indexed. The ingestion spec for the data source is as follows and is saved as metrics-sls.json.
        "type": "sls",
        "dataSchema": {
            "dataSource": "metrics-sls",
            "parser": {
                "type": "string",
                "parseSpec": {
                    "timestampSpec": {
                        "column": "time",
                        "format": "auto"
                    "dimensionsSpec": {
                        "dimensions": ["url", "user"]
                    "format": "json"
            "granularitySpec": {
                "type": "uniform",
                "segmentGranularity": "hour",
                "queryGranularity": "none"
            "metricsSpec": [{
                    "type": "count",
                    "name": "views"
                    "name": "latencyMs",
                    "type": "doubleSum",
                    "fieldName": "latencyMs"
        "ioConfig": {
            "project": <your_project>,
            "logstore": <your_logstore>,
            "consumerProperties": {
                "endpoint": "cn-hangzhou-intranet.log.aliyuncs.com", (In this example, the China (Hangzhou) region is used. Use the intranet endpoint.)
                accessKeyId: "<your-access-key-id>",
                "access-key-secret": <your_access_key_secret>,
                "logtail.collection-mode": "simple"/"other"
            "taskCount": 1,
            "replicas": 1,
            "taskDuration": "PT1H"
        "tuningConfig": {
            "type": "sls",
            "maxRowsInMemory": "100000"
    The ingestion specs of Kafka Indexing Service and LOG Indexing Service are similar. Note the following fields:
    • type: sls
    • dataSchema.parser.parseSpec.format: depends on ioConfig.consumerProperties.logtail.collection-mode (log collection mode of Log Service). If you select Simple Mode, enter the source file format. If you do not select Simple Mode, enter json.
    • ioConfig.project: the project of which the logs you want to collect.
    • ioConfig.logstore: the Logstore of which the logs you want to collect.
    • ioConfig.consumerProperties.endpoint: the intranet endpoint of LOG. For example, the endpoint for the China (Hangzhou) region is cn-hangzhou-intranet.log.aliyuncs.com.
    • ioConfig.consumerProperties.access-key-id: the AccessKey ID of the account.
    • ioConfig.consumerProperties.access-key-secret: the AccessKey Secret of the account.
    • ioConfig.consumerProperties.logtail.collection-mode: the log collection mode of Log Service. If you select Simple Mode, enter simple. Otherwise, enter other.
  2. Run the following command to add a LOG supervisor.
    curl --negotiate -u:druid -b ~/cookies -c ~/cookies -XPOST -H 'Content-Type: application/json' -d @metrics-sls.json http://emr-header-1.cluster-1234:18090/druid/indexer/v1/supervisor
    Notice The --negotiate, -u, -b, and -c options are required for secure Druid clusters.
  3. Import data to LOG

    You can import data to LOG by using multiple methods. For more information, see Log Service.

  4. Perform queries by using Druid.