eG Monitoring
 

Measures reported by StormClustTest

Apache storm cluster is similar to the Hadoop cluster but Storm use topologies instead of MapReduce jobs. In the storm, the topology needs to be terminated otherwise it will run forever unlike MapReduce jobs. Apache storm cluster will have one master node and many worker nodes. Apache Storm uses an internal distributed messaging system for the communication between nimbus and supervisors. Nimbus is a master node of Storm cluster and all other nodes in the cluster are called as worker nodes.

This test reports the number of supervisors, topologies, executors in the Storm cluster and the number of tasks completed by the cluster. Using this test, administrators would be able to identify the number of available CPU cores, the total memory used by the CPU and the available worker slots that are currently existing in the Storm cluster of the target Apache Storm. Any high usage of the CPU memory will alert the administrators to take remedial measures.

Outputs of the test: One set of results for the target Apache Storm.

The measures made by this test are as follows:

Measurement Description Measurement Unit Interpretation
Supervisors Indicates the number of supervisors in the Storm cluster. Number

 
topologies Indicates the number of topologies in the Storm cluster. Number

 
executorsTotal Indicates the number of total executors in the Storm cluster. Number

 
tasksTotal Indicates the number of total tasks executed by the Storm cluster. Number

A very high value is required for this measure.
totalMem Indicates the total size of memory in the Storm cluster. MB

 
availMem Indicates the size of available memory in the Storm cluster. MB

A very high value is required for this measure.
usedMem Indicates the size of used memory in the Storm cluster. MB

A value close to the Total memory measure indicates that the memory resources are depleting rapidly.
availMemPercent Indicates the percentage of available memory in the Storm cluster. Percent

A value close to 100% is required for this measure.
totalCpu Indicates the total number of CPU cores in the Storm cluster. Number

 
availCpu Indicates the number of available CPU cores in the Storm cluster. Number

A very high value is required for this measure.
usedCpu Indicates the number used CPU cores in the Storm cluster. Number

A very low value is required for this measure.
availCpuPercent Indicates the percentage of available CPU from the total CPU. Percent

A value close to 100% is required for this measure.
slotsTotal Indicates the total number of worker slots in the Storm cluster. Number

 
slotsUsed Indicates the number of used worker slots in the Storm cluster. Number

A very high value is required for this measure.
slotsFree Indicates the number of free worker slots in the Storm cluster. Number

A very low value is required for this measure.
slotsUsedPercent Indicates the percentage of used worker slots in the Storm cluster. Percent

A value close to 100% is required for this measure.