api-server

README

api-server

This component is used to send command from UI to Redis, using HTTP Rest API and Redis Pub/Sub protocol.

Building

To build the binary just execute

# get dependencies
go get

# build binary
go build

# build docs
swag init

Configuration

The configuration is read from ENV

LISTEN_ADDRESS=0.0.0.0:8080
REDIS_ADDRESS=127.0.0.1:6379
REDIS_USER=default
REDIS_PASSWORD=""
SECRET=MY_SECRET,11

Running

To execute the binary run

./api-server

You can also specify configurations using env vars

LISTEN_ADDRESS=0.0.0.0:8080 REDIS_ADDRESS=127.0.0.1:6379 REDIS_USER=my-user REDIS_PASSWORD=MyPass SECRET=MY_SECRET,11./api-server

APIs

Each API is authenticated and authorized through a JWT (JSON Web Token). In order to get access to the APIs you have to login before and get a JWT token. SECRET is used to set the JWT signature.

• POST /login

curl -s -H "Content-Type: application/json" -X POST http://localhost:8080/api/login --data '{"username": "<username>", "password": "<password>"}' | jq

{
  "code": 200,
  "expire": "2021-04-21T14:05:35+02:00",
  "token":  "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJhY3Rpb25zIjpbImNyZWF0ZS11c2VycyIsImRlbGV0ZS11c2VycyJdLCJleHAiOjE2MTkwMDY3MzUsImlkIjoiZWRvYXJkbyIsIm9yaWdfaWF0IjoxNjE4NDAxOTM1LCJyb2xlIjoiYWRtaW4ifQ.ru8CbqduPTBI4G9R3zINC_-L38Thggg_9ExFV3Grf18"
}

Once you have the token, each request must include this token in the header.

Authorization: Bearer eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJhY3Rpb25zIjpbImNyZWF0ZS11c2VycyIsImRlbGV0ZS11c2VycyJdLCJleHAiOjE2MTg5MDc2NTIsImlkIjoiZWRvYXJkbyIsIm9yaWdfaWF0IjoxNjE4MzAyODUyLCJyb2xlIjoiYWRtaW4ifQ.dyqSFWi7L0aKAe7mujBJ9eN2nFnC9PcnFPhSQOZc2Nc"

The finally API request should be:

curl -s -H "Authorization: Bearer <your_token>" -X <VERB> http://localhost:8080/api/<api_name> --data '{<your_json_data}' | jq

Basic authentication

There are some cases where it is necessary to validate requests coming from other servers (Caddy, Nginx, ...) using the HTTP Basic authentication method. To achieve this functionality some routes verify authentication using basic auth.

The api-server could check that the action passed from route is among the user (obtained from Basic-Auth credentials) role authorized actions.

Basic authentication API list

• GET /module/:module_id/http-basic/:action Use Basic HTTP auth for module

  Where:

  • :module_id: is the name of the module to check for action
  • :action: is the action to validate in the authorized actions

2FA (Two-Factor authentication)

If you feel comfortable and more secure to use 2FA, you can enable it inside the user setting page o via API. To enable these are the steps:

• Login and get a valid JWT token
• Enable 2FA for a specific user
• Scan auto-generated QR code and add the secret inside your Google Authenticator (or similar protocol app)
• Copy the generated OTP from the app and call the API to verifiy the OTP
• After the OTP verification the JWT token can be used to call APIs

2FA API list

• GET /api/2FA Read current 2FA status

  RESPONSE
  {
    "code": 200,
    "data": true,
    "message": "2FA set for this user"
  }
  
  {
    "code": 200,
    "data": false,
    "message": "2FA not set for this user"
  }
  

• DELETE /api/2FA Revoke 2FA secret

  BODY
  {
    "code": 200,
    "data": "",
    "message": "2FA revocate successfully"
  }
  

• GET /api/2FA/qr-code Get URL with Authenticator setup (can be used to generate a QR Code)

  RESPONSE
  {
    "code": 200,
    "data": {
      "key": "OQPZYGNFP4E37NJUCBR7ELEONC7WSXZW",
      "url": "otpauth://totp/NethServer:admin2fa?algorithm=SHA1\u0026digits=6\u0026issuer=NethServer\u0026period=30\u0026secret=OQPZYGNFP4E37NJUCBR7ELEONC7WSXZW"
    },
    "message": "QR code string"
  }
  

• POST /api/2FA Used to validate the OTP

  BODY
  {
    "otp": "157720",
    "secret": "4W5XXQDHH3ROJBUOHEKUEL6M44IYUOMO"
  }
  
  RESPONSE
  {
   "code": 200,
   "data": null,
   "message": "2FA set successfully"
  }
  

API list

The API doc is generated using swagger. The swagger.json file is built automatically and published at:

https://raw.githubusercontent.com/NethServer/ns8-core/swagdoc/swagger.json

Redis

Each API is mapped to a specific command on Redis:

• GET /api/<cluster|node|module>/tasks ⟶ LRANGE <cluster|node|module>/tasks 0 -1

• POST /api/<cluster|node|module>/tasks ⟶ LPUSH <cluster|node|module>/tasks <payload>

  PAYLOAD contains the task information:

  • id ⟶ task id is an uuid string generated from server
  • action ⟶ action to execute
  • data ⟶ data used from action to execute the task

Audit

Every request made to the server, using its APIs or WebSocket, is logged inside an audit db. The audit db is store in a file using a SQLite database schema. Each record is composed by:

• ID: the unique id of the record, autoincrement
• User: the name of the user that made the specific action
• Action: the name of the action made by the user
• Data: the payload of the action (if present). To avoid storing passwords in clear-text, any attribute in Data matching any of the substrings "password", "secret", "token" is forced to value "XXX".
• Timestamp: the time when the specific action is executed

Websockets

The API server provides a WebSocket connection under the /ws path. Through websocket it is possible to send commands and receive data asynchronously via the socket channel.

Commands

At the moment the supported commands are:

• logs-start: used to get the logs for a specific entity (cluster, node or module). Use loki cli (logcli) to search and retrieve logs
• logs-stop: used to stop the logs you are following (normally used in tail mode)

To execute commands via websocket, you need to send a payload to the websocket and listen to the results.

logs-start

INPUT
{
   "action": "logs-start",
   "payload": {
      "id": "2fd5b630-3d27-4b88-ac11-ad2d09f4bbd7",
      "mode": "dump",
      "lines": "25",
      "filter": "",
      "from": "2021-01-19T10:00:00Z",
      "to": "2021-01-19T20:00:00Z",
      "entity" :"module",
      "entity_name": "traefik1",
      "timezone": "Europe/Rome"
   }
}

• action: fixed to logs-start
• payload: payload of the action
  • id: used to track multiple stream output (used in UI to get different logs stream in different panes)
  • mode: must be tail or dump - string (choose how to retrieve logs)
  • lines: could be empty or a number - string (could be empty in tail mode)
  • filter: could be empty or string - string (used to search specific words inside logs)
  • from: could be empty or iso8601 date string - string (used to search specific logs in a date range)
  • to: could be empty or iso8601 date string - string (used to search specific logs in a date range)
  • entity: must be cluster or node or module - string
  • entity_name: could be empty (cluster case) or name of the entity - string (ex. hostname of the node or module id like traefik1)
  • timezone: could be empty (default UTC) or a specific valid timezone (eg. Europe/Rome)

OUTPUT
{
   "name": "logs-start",
   "payload": {
      "message": "2022-02-17T20:47:24+01:00 fedora33.n2.edoardo --\u003e [GIN-debug] Listening and serving HTTP on 127.0.0.1:8080",
      "pid": "14431"
   },
   "timestamp":"2022-03-01T09:01:41.698943069Z",
   "type":"action"
}

• name: is the name of the action
• payload: contains the response logs
  • message: contains the log message
  • pid: is the pid of the process that actually reads log
• timestamp: timestamp of the action
• type: used to identify the websocket outputs

logs-stop

INPUT
{
   "action": "logs-stop",
   "payload": {
      "id": "2fd5b630-3d27-4b88-ac11-ad2d09f4bbd7",
      "pid": "14088"
   }
}

• action: fixed to logs-start
• payload: payload of the action
• pid: contains the pid of the process that reads log

OUTPUT
{
   "name": "logs-stop",
   "payload": {
      "id": "2fd5b630-3d27-4b88-ac11-ad2d09f4bbd7",
      "pid": "14431",
      "message": "logs follow stopped"
   },
   "timestamp":"2022-03-01T09:01:41.698943069Z",
   "type":"action"
}

• name: is the name of the action
• payload: contains the response logs
  • id: id of the trackes log stream
  • pid: is the pid of the process that actually reads log
• timestamp: timestamp of the action
• type: used to identify the websocket outputs

Task events

The websocket also listens for every event of every task launched within the cluster and reports the payload through the socket channel.

api-server-logs

This component is a wrapper of api-server websocket logs

Usage

# ./api-server-logs

api-server-logs is a wrapper of api-server websocket logs

Usage:
  api-server-logs [command]

Available Commands:
  completion  Generate the autocompletion script for the specified shell
  help        Help about any command
  logs        get logs of a specific entity. default: cluster
  version     show api-server-logs version

Flags:
  -h, --help   help for api-server-logs

Use "api-server-logs [command] --help" for more information about a command.

Commands

• completion:

  Generate the autocompletion script for api-server-logs for the specified shell.
  See each sub-command's help for details on how to use the generated script.
  
  Usage:
    api-server-logs completion [command]
  
  Available Commands:
    bash        Generate the autocompletion script for bash
    fish        Generate the autocompletion script for fish
    powershell  Generate the autocompletion script for powershell
    zsh         Generate the autocompletion script for zsh
  
  Flags:
    -h, --help   help for completion
  

• help: prints the Usage output

• logs: the command to retrieve logs in dump or follow mode

  get logs of a specific entity. default: cluster
  
  Usage:
    api-server-logs logs [flags]
  
  Flags:
    -e, --entity string   get logs for a specific entity: cluster, node, module (default "cluster")
    -f, --from string     get logs from a specific date. ISO8601 format
    -h, --help            help for logs
    -l, --lines string    get logs for a specific lines in dump mode (default "25")
    -m, --mode string     get logs in a specific mode: tail or dump (default "tail")
    -n, --name string     get logs for a specific entity name. used in node or module
    -s, --search string   get logs for a specific search string
    -t, --to string       get logs to a specific date. ISO8601 format
    -z, --timezone string get logs in a specific timezone
  

• version: prints the command version

  api-server-logs 0.0.1