Health checks

The gateway service determines if a worker still can be used safely by frequently checking its health. Health checks are used for health determination. Jitar has no out-of-the-box health checks, but you can create and add your own. A common use case is for checking database availability. In case a worker can't access its database, we want the gateway to stop using it. Health checks can be added to any of the services.

BREAKING CHANGES

Version 0.5 introduced breaking changes. Please check our migration guide for more information.

In this section you'll learn how to create and add your own health checks.

Creating health checks

Jitar provides an interface you have to implement in order to add a custom check to a service. Let's look at a simple example:

// src/DatabaseHealthCheck.ts
import { HealthCheck } from 'jitar';

export default class DatabaseHealthCheck implements HealthCheck
{
    constructor(/* ... */) { /* ... */ }

    get name() { return 'database'; } // New required property!

    get timeout() { return undefined; }
    
    async isHealthy(): Promise<boolean> { /* ... */ }
}

The health check interface requires you to implement the name and timeout getter. The name is used for its registration, and needs to be unique. The timeout specifies the time in milliseconds that the service waits before checking for the health again. If a timeout is not required, return undefined. Otherwise set the time in milliseconds.

Adding health checks

Health check needs to be registered at the service configuration. Only health check instances can be registered. Each instance needs to be exported as default in a separate module file like this:

// src/defaultDatabaseHealthCheck.ts
import DatabaseHealthCheck from './DatabaseHealthCheck';

const instance = new DatabaseHealthCheck(/* ... */);
export default instance;

We can use this module file for the registration at the service:

json

// services/worker.json
{
    "url": "http://localhost:3000",
    "healthChecks": ["./databaseHealthCheck"],
    "worker":
    {
        // ...
    }
}

Once added, the worker will trigger the check automatically. You can also check yourself using the health API. More information on this can be found in the MONITOR section.

Note that health checks are defined at the root level of the configuration. This means you can add health checks for all service types. When using external monitoring tools that monitor a cluster, this could be useful.

Health checks ​

Creating health checks ​

Adding health checks ​

Health checks

Creating health checks

Adding health checks