Skip to content

Latest commit

 

History

History
79 lines (56 loc) · 4.78 KB

how.md

File metadata and controls

79 lines (56 loc) · 4.78 KB
title summary reviewed
How does ServiceControl work?
An overview of how ServiceControl collects and processes messages and data from an NServiceBus system
2022-10-19

ServiceControl is a background process that will collect and store data and make it available via an HTTP API to ServicePulse and ServiceControl.

Note

The ServiceControl HTTP API may change at any time. It is designed for use by ServicePulse and ServiceInsight only. The use of this HTTP API for other purposes is not supported.

How ServiceControl receives data

For each ServiceControl service type (i.e. error, audit, and monitoring) there is different data available that must be provided by user-built NServiceBus endpoints.

graph LR
  subgraph Endpoints
    Error
    Plugins[Saga audit<br>Heartbeats<br>Custom checks]
    Audit
    Monitoring
  end

  Audit -- Audit<br>data --> AuditQ(Audit queue)
  AuditQ --> ServiceControlAudit[ServiceControl<br>audit instance]
  ServiceControlAudit --> AuditLog(Audit.Log<br>queue)
  ServiceControlAudit -. Integration<br>events .-> Watchers

  Plugins -- Plugin data --> SCQ
  Plugins -- SagaAudit<br>data --> AuditQ(Audit queue)
  SCQ(ServiceControl<br>queue) --> ServiceControlError

  Error -- Error<br>data --> ErrorQ(Error queue)
  ErrorQ --> ServiceControlError[ServiceControl<br>error instance]
  ServiceControlError --> ErrorLog(Error.Log<br>queue)
  ServiceControlError -. Integration<br>events .-> Watchers[Subscribers]

  Monitoring --> MonQ(Monitoring queue)
  MonQ --> ServiceControlMonitoring[Monitoring instance]

  classDef Endpoints fill:#00A3C4,stroke:#00729C,color:#FFFFFF
  classDef ServiceControlInstance fill:#A84198,stroke:#92117E,color:#FFFFFF,stroke-width:4px
  classDef Plugin fill:#00729C,stroke:#000,color:#ddd

  class ServiceControlError ServiceControlInstance
  class ServiceControlAudit ServiceControlInstance
  class ServiceControlMonitoring ServiceControlInstance
  class Watchers Endpoints
  class Plugins Plugin
  class Audit Plugin
  class Error Plugin
  class Monitoring Plugin
Loading

Error instances

Recoverability is an important feature in NServiceBus. It enables automatic retries and continuity within a system, as failed messages will be moved aside to allow other messages to be processed while the errors are investigated. Those error messages contain business data that must eventually be processed.

NServiceBus will move messages it cannot process to an error queue. This is where ServiceControl comes into play to consume these messages. ServiceControl will pick up the message and store it in an internal database. ServiceControl uses an embedded RavenDB instance for the internal database. After it is stored in the database, the message is made available to ServicePulse and ServiceInsight for visualization, retries, and other operations.

Note

It is recommended not to provide end-users with the ability to retry messages. The message could fail again and end up in ServiceControl once again. It could be even more problematic when many messages are retried during a peak in message processing. This will result in even more messages being processed by an endpoint, causing valid messages to be delayed even longer.

Find out more about failed messages in ServicePulse.

Audit instances

To enable ServiceInsight to visualize the message flow through the system, it must have access to every message that has been successfully processed by the system. This requires endpoints to enable auditing. ServiceControl consumes these messages and stores them in its internal database.

ServiceInsight will retrieve the data from ServiceControl via the HTTP API and use header information (added by NServiceBus during message processing) to figure out which message caused other messages to be sent, including which sagas were accessed when the SagaAudit plugin is configured in an endpoint.

Monitoring instances

For ServicePulse to report metrics on logical endpoints, endpoint instances and on specific messages, each endpoints needs to be enabled to forward metrics to ServiceControl. ServicePulse will then be able to retrieve the data from ServiceControl via the HTTP API.

Forwarding

ServiceControl consumes messages from the audit and error queues. That is, it removes all messages from those queues. If a copy of those messages is required for further processing, configure audit forwarding and/or error queue forwarding.