Processors

Processors are the core of the Joule platform, each performing a specific task. They create use case when linked together

Overview

The Joule platform uses processors which are chained to form pipelines.

It consists of one or more linked processors, each performing a specific task based on functional programming principles.

Events are processed in sequence, with the final event either sent to a specific processor or directly to the connected systems. Enabling business use cases, while providing metrics for observability via JMX monitoring.

A pipeline that forms a use case

---

Types of Processors

Joule offers several core processors that developers can use to build different use cases.

These processors are categorised by function, based on their role within the platform.

Filters

Reduce stream processing overhead by filtering irrelevant events

Enrichment

Enrich events using linked reference data, metrics and analytics

Transformation

Transform event attribute values based upon a desired target state

Triggers

Apply rule based triggers for downstream business actions

---

Observability and Metrics

Each OOTB processor record and presents event metrics to JMX monitoring platforms. This provides direct insight into how a streaming use case is performing.

Metrics available

• Received

• Processed

• Failed

• Discarded

• Ignored

• Average processing latency

All of these metrics are enabled by default and presented as JMX beans on the standard monitoring ports.

Enable JMX

Apply these settings to the joule execution path

-Dcom.sun.management.jmxremote.port=5000
-Dcom.sun.management.jmxremote.rmi.port=5000

For further information use the instructions in Monitoring and Management Using JMX Technology

Core attributes

Every processor that inherits from the base class AbstractProcessor , using the Joule SDK, will gain these core features.

Attribute	Description	Data Type	Required
name	Uniquie name of processor.	String Default: Random UUID
enabled	Toogle to enable processor processing	Boolean Default: true
properties	Processor specific properties	Map
clone event	Toogle to clone event for processing. Note this will cause significant memory overhead for high load use cases	Boolean Default: false
stores	External data stores can be attached, useful for analytics, enrichment, filtering and transformation use cases
latency trigger	Number of events processed to trigger the average processing calculation. A value less than 1 switches off the calculation.	Integer Default: 1000