This article is part of the NServiceBus Learning Path.
The Custom Checks plugin enables endpoint health monitoring by running custom code and reporting status (success or failure) to a ServiceControl instance.
graph LR subgraph Endpoint CustomChecks[Custom Checks] end CustomChecks -- Custom Check<br>Data --> SCQ[ServiceControl<br>Input Queue] SCQ --> SC[ServiceControl<br>instance] SC -. Integration<br/>Events .-> Integration[Integration<br/>Event Handler] SC -- Custom Check<br>Data --> ServicePulse
Set up custom checks
To enable custom checks in an environment:
- Install a ServiceControl instance
- Install the Custom Checks plugin in endpoints that will contain custom checks and write a custom check
- View the status of custom checks in ServicePulse
- Optionally subscribe to integration events from ServiceControl when custom checks succeed or fail
Common scenarios for custom checks
While the specific needs and dependencies of solutions and endpoints may vary significantly, there are common scenarios of required external conditions and services.
Most endpoints require connectivity to local intranet or external internet. Some also require VPN connectivity.
It is recommended to periodically check that the endpoint (and its host) can indeed connect to the resources it requires.
- In a store & forward pattern (MSMQ): are messages getting stuck in outgoing queues because of a connectivity problem between two machines?
- In a broker pattern: can the endpoint connect to the broker (e.g., SQL Server)?
- If required by the endpoint, can the endpoint connect to the local intranet?
- If required by the endpoint, can the endpoint connect to the internet?
- If required by the endpoint, can the endpoint connect with the required security settings, credentials, and VPN software?
- When local or remote storage is required by the endpoint, is that storage location available, accessible, and properly configured (security, permissions, quota, etc.)?
- When local or remote storage is required by the endpoint, is there enough available storage left for required operations?
- If there is a requirement for a minimum free space available, is that requirement met?
- Gather a collection of measurements over time and extrapolate when the storage will run out. Is that time below some critical threshold?
- When an endpoint needs to communicate with external services (credit card validation, identity and authentication service, CRM, etc.), is that service available, accessible, and responsive?