JVMThreadStatesRecorder

Purpose of the JVMThreadStatesRecorder is to write JVM thread states (like runnable, blocked ...) to the InfluxDB and then visualize those states in the Grafana, etc. 🧵

Command line (InfluxDB parameters are optional, could be configured with /influxdb-configuration):

java -jar JVMThreadStatesRecorder.jar 7080 http://127.0.0.1:8086 JVMThreadStatesRecorder

1. java (version 11 or later)
2. JVMThreadStatesRecorder.jar can be taken from out/artifacts/JVMThreadStatesRecorder_jar/
3. 7080 (a port to start web server; a default port: 7070)
4. http://127.0.0.1:8086 (an url to the InfluxDB; optional)
5. JVMThreadStatesRecorder (DB name in the Influxdb; optional)

Endpoints

• Configure a connection to the InfluxDB

POST /influxdb-configuration
body:

{  
"influxdbUrl":"http://127.0.0.1:8086",  
"influxdbDb":"JVMThreadStatesRecorder",  
"influxdbMeasurement":"thread_states",  
"influxdbBatchSize":"1000",  
"influxdbBatchTime":"10"  
}

"influxdbDb":"jvm_app_monitoring", (optional, default: JVMThreadStatesRecorder)
"influxdbMeasurement":"thread_states", (optional, default: thread_states)
"influxdbBatchSize":"1000", (optional, default: 1000)
"influxdbBatchTime":"10" (optional, default: 10)

• Start monitoring an application with a jmx host and a jmx port

POST /start
body:

{  
"jmxHost":"127.0.0.1",  
"jmxPort":"9010",  
"threadFilter":"Example",
"threadStates":"true",   
"contentionMonitoring":"true",  
"tags":{  
"tag1":"value1",  
"tag2":"value2"  
}  
}  

response: id of the task that can be used for stopping the task
"threadFilter":"Example" (optional, default: all threads, carefull, there could be hundreds of threads) "threadStates":"true" (true or false. Enable thread states monitoring)
"contentionMonitoring":"true" (true or false. Enable contention monitoring, blocked count and blocked time)
"tags":{....} (optional, additional tags for enriching data in the InfluxDB)

• Start monitoring an application with a pid

POST /start
body:

{  
"pid":"44620",  
"threadFilter":"Example",
"threadStates":"true",
"contentionMonitoring":"true", 
"tags":{  
"tag1":"value1",  
"tag2":"value2"  
}  
}  

response: id of the task that can be used for stopping the task
"threadFilter":"Example" (optional, default: all threads, carefull, there could be hundreds of threads)
"threadStates":"true" (true or false. Enable thread states monitoring)
"contentionMonitoring":"true" (true or false. Enable contention monitoring, blocked count and blocked time)
"tags":{....} (optional, additional tags for enriching data in the InfluxDB)

• Stop the task with an id

GET /stop/id={id of the task}

• List of currently running tasks

GET /tasks
response:

{  
  (id of the task):  
  {  
    (Configuration.Configuration object of the task)  
  },  
  ...
}    

• Returns connection configuration to the InfluxDB

GET /dbconfig

• Start internal monitoring

GET /internal-monitoring
Shows time in ms. which is spent on collecting thread states, grouped by a task

Grafana configuration

• Ready to use dashboard

📈 JVM_Thread_States_Dashboard.json
Tested with Grafana 8.4.3

• Visaulization: State timeline

• Possible query

• Value mappings

Contention monitoring

Shows total number of times that a thread blocked to enter or reenter a monitor and time that a thread has blocked to enter or reenter a monitor
📈 JVM_Thread_States_Dashboard.json

Internal monitoring

Shows time in ms. which is spent on collecting thread states, grouped by a task
📈 JVM_Thread_States_Dashboard.json