Skip to content

Druid

Collect metrics related to the Druid database connection pool.

Installation and Configuration¶

Expose metrics related to the Druid database connection pool via JMX and read them using DDTrace. DDTrace reports data through its internal statsd module.

1. Download DDTrace¶

wget -O dd-java-agent.jar 'https://static.guance.com/dd-image/dd-java-agent.jar'

2. Configure Druid JMX¶

Create a file named druid.yaml with the following content:

init_config:
  is_jmx: true
  collect_default_metrics: true

instances:
  - jvm_direct: true
    name: duird-DruidDataSource-monitoring
    collect_default_jvm_metrics: false
    collect_default_metrics: false
    refresh_beans: 60
    conf:
      - include:
          bean_regex: "com.alibaba.druid:type=DruidDataSource,id=\\d+"
          tags:
            pool: druid
          attribute:
            MaxActive:
              metric_type: gauge
              alias: druid.max_active
            ConnectCount:
              metric_type: gauge
              alias: druid.connect_count
            WaitThreadCount:
              metric_type: gauge
              alias: druid.wait_thread_count
            ActivePeak:
              metric_type: gauge
              alias: druid.active_peak
            InitialSize:
              metric_type: gauge
              alias: druid.initial_size
            ConnectErrorCount:
              metric_type: gauge
              alias: druid.connect_error_count
            ActiveCount:
              metric_type: gauge
              alias: druid.active_count
            CloseCount:
              metric_type: gauge
              alias: druid.close_count
            PoolingCount:
              metric_type: gauge
              alias: druid.pooling_count
            RecycleCount:
              metric_type: gauge
              alias: druid.recycle_count
            CreateCount:
              metric_type: gauge
              alias: druid.create_count
            DestroyCount:
              metric_type: gauge
              alias: druid.destroy_count
            CommitCount:
              metric_type: gauge
              alias: druid.commit_count
            RollbackCount:
              metric_type: gauge
              alias: druid.rollback_count              
            MaxWait:
              metric_type: gauge
              alias: druid.max_wait
            MinIdle:
              metric_type: gauge
              alias: druid.min_idle
            MaxIdle:
              metric_type: gauge
              alias: druid.max_idle

3. Adjust the startup command¶

java \
-javaagent:/xxx/dd-java-agent.jar \
-Ddd.agent.port=9529 \
-Ddd.service=<application-name> \
-Ddd.jmxfetch.check-period=1000 \
-Ddd.jmxfetch.enabled=true \
-Ddd.jmxfetch.config.dir=/xxx/ \
-Ddd.jmxfetch.config=druid.yaml \
-jar xxxx.jar

4. DataKit Collector Configuration¶

Since Druid metrics are mainly processed and reported through DDTrace, enabling DDTrace will simultaneously collect APM and metrics data. Therefore, two collectors need to be enabled.

Enable the ddtrace collector

cd /usr/local/datakit/conf.d/ddtrace cp ddtrace.conf.sample ddtrace.conf

Enable the statsd collector

cd /usr/local/datakit/conf.d/statsd cp statsd.conf.sample statsd.conf

Restart DataKit

systemctl restart datakit

Metrics¶

Metric Set `druid`¶

Metric	Description	Unit
active_count	The number of connections currently in use.	count
active_peak	The historical peak of active connections.	count
close_count	The number of closed connections.	count
commit_count	The number of committed transactions.	count
connect_count	The total number of successfully established connections.	count
connect_error_count	The number of failed connection attempts.	count
create_count	The number of created connections.	count
destroy_count	The number of destroyed connections.	count
initial_size	The number of connections created when initializing the connection pool.	count
max_active	The maximum number of active connections allowed in the connection pool.	count
max_idle	The maximum number of idle connections allowed in the connection pool.	count
max_wait	The maximum wait time for obtaining a connection from the connection pool.	ms
min_idle	The minimum number of idle connections allowed in the connection pool.	count
pooling_count	The number of currently idle connections.	count
recycle_count	The number of recycled connections.	count
rollback_count	The number of rolled-back transactions.	count
wait_thread_count	The number of threads currently waiting to acquire a connection.	count