Loads configuration from JSON files from a <app_root>/config directory. The NODE_ENV environment variable determines which configuration file is loaded. Settings that are shared between environments can be put in the optional default.json. Variables loaded from the environment files take precedence over the default.
It's also possible to override configuration values using a file named .env in <app_root> and by specifying them as environment variables.
You should use this module instead of using if/switch statements and the NODE_ENV environment variable directly. This will make your application easier to configure when it grows.
For node versions below 4 please use v1.3.4 - npm i -S [email protected].
npm install exp-config
Create a file named development.json in a folder named config in the application's root directory, such as:
{
"someProp": "some value"
}In your code require exp-config and retrieve the configuration value:
const config = require("exp-config");
const configuredValue = config.someProp;You can also nest properties in the configuration files:
{
"server": {
"host": "google.com"
}
}const config = require("exp-config");
const configuredValue = config.server.host;Booleans may need special care:
const config = require("exp-config");
if (config.boolean("flags.someFlag")) {
...
}This is to prevent config.flags.someFlag having the value "false" (which is truthy) to cause problems.
By default exp-config loads <app_root>/config/development.json. This behavior is typically used for local development and changed by specifying a different environment using the NODE_ENV environment variable, like this:
$ NODE_ENV=production node app
When starting an application in this way exp-config will instead load <app_root>/config/production.json. Likewise, it's common to have a separate configuration file for tests, and use NODE_ENV=test when running them.
In some cases you want or need to use NODE_ENV=production to get for instance performance benfits in Express.js. But we still want load specific configuration for an evironment you can use NODE_CONFIG_ENV to use for configuration identification.
NODE_ENV=production NODE_CONFIG_ENV=qa node app
The qa configuration would be used instead of the production configuration.
Individual values in the loaded configuration can be overridden by placing a file named .env in the application's root (<app_root>/.env). An example .env file can look like this:
someProp=some other value
server.host=example.com
flags.someFlag=true
# The .env file can contain comments which is nice
# when you want to easily switch between values
#server.host=prod.example.com
#server.host=stage.example.com
#server.host=test.example.com
If you use nodemon to automatically restart your app while developing, you should add "watch": ["*", ".env"] to your nodemon.json file so that the app is restarted whenever you change your .env file.
It's also possible to override configuration by specifying them as environment variables when starting the application, like this:
$ someProp=value node app
To override nested properties with environment variables do like this:
$ env 'flags.someFlag=false' node .
By default exp-config uses a file called .env in the root folder, you can override this by setting an environment variable named ENV_PATH to the new files path and name. NOTE: this is relative to the projects root folder.
$ ENV_PATH=relative/env/path/.envfile node /home/someuser/myapp/app.js
Values are loaded with the following precedence:
- Environment variable
- .env file
- Configuration file
In other words, environment variables take precedence over .env files and configuration files.
NOTE, there is one exception: When NODE_ENV equals test (NODE_ENV=test) the .env file and environment variables are ignored. We want the test process to be as isolated and repeatable as possible, and are therefore minimizing the possibility of sticky human fingers messing with its configuration.
NOTE II, exception to the exception: If you want environment variables to be honored in the test environment, you can set the ALLOW_TEST_ENV_OVERRIDE environment variable. This is useful for overriding certain configurations when doing in-container testing. The .env file will still be ignored however.
In openshift and some versions of alpine, you are not allowed to set environvariables with periods (".") in them. To solve this exp-config allows you to set INTERPRET_CHAR_AS_DOT to any char you like to be interpret as a period. Setting INTERPRET_CHAR_AS_DOT=_ and foo_baz_bar="value" will set the value foo.baz.bar to "value" as long as foo.baz.bar it exists in the config-file.
By default exp-config tries to locate the config folder and the (optional) .env file by using process.cwd(). This works great when starting the application from it's root folder. However, sometimes that's not possible. In such cases the root path can be specified by setting an environment variable named CONFIG_BASE_PATH, like this:
$ CONFIG_BASE_PATH=/home/someuser/myapp/ node /home/someuser/myapp/app.js
By default exp-config allows the values to be overriden by bash variables. Setting the ENV_PREFIX enables you to override your variables even if they are injected with a prefix. For example, if your environment makes s3 settings accessible for you via the bash variables S3_key, S3_secret and S3_bucket, and you want to override the settings for key, secret and bucket in your config file, setting ENV_PREFIX=S3_ allows you to do this.
$ ENV_PREFIX=S3_ node /home/someuser/myapp/app.js
An application using exp-config typically have a directory structure like this:
.
├── .env <-- Overrides for local development, not committed to source control
├── config <-- Configuration files committed to source control
| ├── development.json <-- used during local development, loaded if NODE_ENV is unset
| ├── production.json <-- used in production by setting NODE_ENV
| └── test.json <-- used in tests by setting NODE_ENV
| └── default.json <-- shared settings, optional
└── app.js <-- the app