npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2024 – Pkg Stats / Ryan Hefner

@qatalog/gcp-config

v1.0.0

Published

GCP-secrets integrated config loader

Downloads

7,429

Readme

@qatalog/gcp-config

What's this?

A simple config loader for Node.js, integrated with @google-cloud/secret-manager so you can load config settings directly from GCP without leaking them to the environment.

What's wrong with environment variables?

Environment variables can be read by anyone with access to a virtual machine or container. If you set secrets using environment variables, anyone with access to your production containers also has access to your production secrets. Often you don't want that.

How do I use it?

Call the load method with your GCP project id and a config schema that dictates how settings are loaded:

const gcpConfig = require('@qatalog/gcp-config');
const joi = require('joi');

const CONFIG_SCHEMA = {
  foo: {
    secret: 'name_of_secret_in_gcp',
  },

  bar: {
    default: 'default value',
    secret: 'name_of_another_secret_in_gcp',
  },

  // ...
};

main();

async function main() {
  const config = await gcpConfig.load({
    project: process.env.GCP_PROJECT,
    schema: CONFIG_SCHEMA,
  });

  assert(typeof config.foo === 'string');
  assert(typeof config.bar === 'string');

  // ...
}

Can schema properties be nested?

Yes. Schema properties can be nested to arbitrary depth, generating the equivalent tree structure in the returned config object:

const config = await gcpConfig.load({
  project: process.env.GCP_PROJECT,

  schema: {
    foo: {
      bar: {
        secret: 'foo_bar',
      },

      baz: {
        qux: {
          secret: 'foo_baz_qux',
        },
      },
    },
  },
});

assert(typeof config.foo === 'object');
assert(typeof config.foo.bar === 'string');
assert(typeof config.foo.bar.baz === 'object');
assert(typeof config.foo.bar.baz.qux === 'string');

Can I still read non-secret properties from the environment?

Yes. Use the env property in your schema to load a value from an environment variable rather than GCP Secret Manager:

const config = await gcpConfig.load({
  project: process.env.GCP_PROJECT,

  schema: {
    foo: {
      env: 'FOO',
    },

    bar: {
      secret: 'bar',
    },

    baz: {
      env: 'BAZ',
      secret: 'baz',
    },
  },
});

It's fine to specify both the env and secret properties on a single node. When both are set, environment variables take precedence over secrets. This can be especially useful if you want to test changes locally, without touching shared secrets.

Can I read non-secret properties from file?

Yes. Pass the file option to load:

const config = await gcpConfig.load({
  file: path.join(__dirname, `${process.env.NODE_ENV}.json`),

  project: process.env.GCP_PROJECT,

  schema: {
    foo: {
      secret: 'foo',
    },

    bar: {
      baz: {
        secret: 'bar_baz',
      },

      qux: {
        secret: 'bar_qux',
      },
    },
  },
});

file should be the path to a JSON file that matches the structure of your schema:

{
  "foo": "wibble",
  "bar": {
    "qux": "blee"
  }
}

The file does not need to contain every property from the schema and any additional properties will be ignored.

Properties loaded with the file option take precedence over default values, but properties loaded via secret or env take precedence over file.

Can I specify validation options in the schema?

Yes. Each node can have its own schema property that's used for validation and type coercion. We use joi for validation so you can set schema to any joi schema:

const joi = require('joi');

const config = await gcpConfig.load({
  project: process.env.GCP_PROJECT,

  schema: {
    foo: {
      schema: joi.valid('wibble', 'blee'),
      secret: 'foo',
    },

    bar: {
      schema: joi.number().integer().positive(),
      secret: 'bar',
    },
  },
});

assert(typeof config.foo === 'string');
assert(typeof config.bar === 'number');

If a config value violates its schema, the promise returned by load will be rejected with the relevant error from joi.attempt(). This is true regardless of whether the value was loaded using env, secret, file or default.

See the joi api docs for more information about validation and type coercion.

Can I specify type coercion options in the schema?

Yes.

  • Nodes with a joi.array() schema can have a coerce option that marshalls the JSON array string to a JS array object:

    const config = await gcpConfig.load({
      project: process.env.GCP_PROJECT,
    
      schema: {
        foo: {
          coerce: {
            from: 'string',
            to: 'array',
          },
          schema: joi.array().items(joi.valid('wibble', 'blee')),
          secret: 'foo',
        },
      },
    });
    
    assert(Array.isArray(config.foo));
  • Nodes with a joi.string().isoDuration() schema can have a coerce option that marshalls the duration string to the equivalent number of milliseconds or seconds:

    const config = await gcpConfig.load({
      project: process.env.GCP_PROJECT,
    
      schema: {
        foo: {
          coerce: {
            from: 'duration',
            to: 'milliseconds',
          },
          schema: joi.string().isoDuration(),
          secret: 'foo',
        },
      },
    });
    
    assert(typeof config.foo === 'number');

It's likely we'll add more type coercion options in future.

Can I flag properties as required?

Yes. Set required: true on those nodes in your schema:

const config = await gcpConfig.load({
  project: process.env.GCP_PROJECT,

  schema: {
    foo: {
      required: true,
      secret: 'foo',
    },

    bar: {
      required: true,
      secret: 'bar',
    },
  },
});

If required properties are not set by any data source, the promise returned from load will be rejected. Required properties with default values will never fail.

What is the full list of properties I can set in the schema?

All properties are optional:

  • env: Environment variable to read the value from. Environment variables take precedence over secrets.

  • secret: The key to lookup the value in GCP Secret Manager.

  • default: The default value, used as fallback if no other values are found in the environment, Secret Manager or from file.

  • schema: Joi validation schema for the value.

  • coerce: Instructions for type coercion. Value is an object with two properties, from and to. Only supports coercion from ISO 8601 duration strings to numeric milliseconds or seconds right now.

  • required: Boolean indicating whether absence of the value should be treated as an error. Defaults to false.

What happens with secrets that are disabled?

Disabled secrets are ignored, effectively they don't exist at all.

How are secrets with multiple versions handled?

The greatest, non-disabled version of the secret will be used by default. For using a specific version set version: <your_version> in the schema.

const config = await gcpConfig.load({
  project: process.env.GCP_PROJECT,

  schema: {
    foo: {
      version: 1, // Specific version. Make sure this version exists and is not disabled.
      secret: 'foo',
    },
  },
});

Can secret names be prefixed with the runtime environment?

Yes. If you prefix the keys for your secrets in GCP, e.g. production_foo and integration_foo, you can specify that prefix in the call to load and keep the body of your schema constant across all environments:

const config = await gcpConfig.load({
  prefix: `${process.env.NODE_ENV}_`,

  project: process.env.GCP_PROJECT,

  schema: {
    foo: {
      // Might load `production_foo` or `development_foo` etc
      secret: 'foo',
    },
  },
});

Is there a way to disable reading secrets at runtime?

Yes. Secrets will not be loaded if the ignoreSecrets option to load is true:

const config = await gcpConfig.load({
  ignoreSecrets: process.env.NODE_ENV === 'test',

  project: process.env.GCP_PROJECT,

  schema: {
    foo: {
      default: 'default value',
      // Won't be loaded when `NODE_ENV` is `test`
      secret: 'foo',
    },
  },
});

This can be useful in test environments, when running many tests in quick succession can exceed the available read quota for Secret Manager. Nobody likes flaky tests.

How do I run the tests?

For the tests to pass locally, you must be authenticated with gcloud:

gcloud auth login

You also need to set the GCP_PROJECT environment variable. The tests will create, read and delete secrets, so you should use a clean project that doesn't contain any real infrastructure. Your IAM user or service account will need the Secret Manager Admin and Secret Manager Secret Accessor roles.

With all that set, you can run the tests with npm t:

GCP_PROJECT=my-gcp-project npm t

What versions of Node does it support?

Minimum supported node version is 18.

Is there a changelog?

Yes.

What license is it released under?

MIT.