A microservice to perform health checks on turn servers.
The following environment variables can be set:
The port of the microservice. Defaults to 3000.
Can be used to restrict the microservice to specific turnservers. If not set, it will allow health-checks for any
turnserver. Multiple turn-servers must be separated by ; like so:
turn:domain.tld:3478?transport=udp;turn:example.com:1234
The used log level. Accepts the following values: error, warn, info, debug, and trace. Defaults to info.
A health-check can be performed with a simple HTTTP GET:
curl -X GET 'http://localhost:3000?url=turn:domain.tld:3478%3Ftransport=udp&username=example&credential=example'
It accepts the following query parameters:
| Parameter | Required | Description |
|---|---|---|
url |
yes | The turn server url |
username |
no | The username, used for authentication |
credential |
no | The credential, used for authentication |
format |
no | The format of the response. Accepts json, flag, or http-status. Defaults to json |
The answer of the health check depends on the chosen format:
| Format | Response |
|---|---|
json |
The body is a json object with a boolean field available. The HTTP status is always 200 |
flag |
The body is 1 if the turn-server is online, otherwise 0. The HTTP status is always 200 |
http-status |
Always an empty body. The HTTP status is 200 if the turn server is available, otherwise 418 |
The only REST-conform response code for a failed health-check is 200, as the request itself was successful, only the health-check did fail.
It is wrong to send a 4xx CLIENT ERROR (or even worse 5xx SERVER ERROR), as it is in fact no client error. However, many health-checks (e.g. for
Kubernetes) do not look at the response body, but only at the http status code. For these clients, you can use the
http-status format.
yarn dev will build the project and start it afterwards.
Alternatively, you can also run yarn build, followed by yarn start.