Collector Type: Agent

Category: Application Monitors

Application Name: Spark

Global Template Name: Linux - Apache Spark Monitors

Introduction

Apache Spark is an open-source distributed general-purpose cluster-computing framework. Spark provides an interface for programming entire clusters with implicit data parallelism and fault tolerance.

Pre-requisite: Spark REST API

Parameters

NameDefault ValueDescription
ProtocolhttpThe protocol to be used.
Host IP Address127.0.0.1The host on which Spark is running.
Spark Web UI Port8080The port on which Spark Web UI is running.

Note: All field attributes are mandatory, use default values wherever applicable.

Collected Metrics

Metric NameDisplay NameDescription
spark.workersSpark WorkersThe number of workers connected to the master.
spark.coresSpark CoresThe number of CPUs available for all workers.
spark.cores.usedSpark Cores UsedThe number of CPUs used for all applications.
spark.applications.activeSpark Applications ActiveThe number of applications waiting or running.
spark.applications.completedSpark Applications CompletedThe number of application completed.
spark.drivers.activeSpark Drivers ActiveThe number of drivers available.
spark.statusSpark StatusAvailable status of spark master ex:alive.
spark.memorySpark MemoryCalculates the total memory available on spark master.
spark.memory.usedSpark Memory UsedCalculates the memory used by the applications on spark master.