Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

https://github.com/ipfabric/ipfabric-slack-integration


https://github.com/ipfabric/ipfabric-slack-integration

Last synced: 3 months ago
JSON representation

Lists

README 

        
# THIS REPO IS NOW ARCHIVED!

HEAD TO https://github.com/community-fabric FOR UPDATED REPOS OR https://github.com/community-fabric/awesome-ipfabric-list FOR IP FABRIC RELATED LINKS



## Webhook Worker / testing webhook agent



In this directory you can find a simple webhook agent (receiver) that you can use for testing the webhook sub-system

of the IP Fabric platform. In version 1.1 we added support for Slack chat-bot, so commands can now be triggered

even by the user from Slack.



To start it, you need to:

```

$ yarn install

$ yarn start [config/config.json]

```



(In the following text, the `highlighted` words are fields from config.json file.)



The agent will start HTTP server, listening on `port`.



### Webhook notifications processing



When a HTTP POST message arrives on `routeWebhook` route, the agent prints information about it

and reads raw payload attached to the message.



Then it calculates SHA256 HMAC hash of the payload with provided `secret` and compares it with

server-calculated hash of the payload sent in X-IPF-Signature HTTP header. In case the hashes

don't match, the message is dropped and "400 Bad request" is sent back. Note: the `secret` string

is the one you configured in IPF UI while creating the webhook.



Otherwise the payload is legitimate and it is parsed (if parsing fails, "400 Bad request" is sent

back as well). The parsed payload is printed to the output and the agent sends "200 OK" back

to indicate confirmation of successfully processed webhook. Then the payload is passed to the

payload handler that, in this demo, examines the payload, fetches extra data from IP Fabric platform

as needed, and posts notifications to Slack.



### Slack chatbot commands processing



When a HTTP POST message arrives on `routeChatbot` route, the agent prints information about it

and reads, parses, and verifies raw payload attached to the message.



In case the payload corresponds to a supported Slack command, the message is acknowleged to Slack

and the payload is passed to chatbot handler. Otherwise the message is also acknowleged to Slack,

but the response says the command is missing or not supported or has wrong arguments, and it lists

the supported commands and their arguments.



The chatbot handler fetches data from IP Fabric platform as needed and posts processed results to Slack.



### Slack integration: webhooks



To integrate with Slack, you need an OAuth access token that you can use to authenticate

and send messages to given Slack workspace / channel(s).



In case you don't have one yet, navigate to [https://api.slack.com/apps](https://api.slack.com/apps)

and create new application there. In Features / OAuth & Permissions, you need to add some Bot

Token Scopes, depending on what features your application will be using:

* `chat:write` is for sending messages,

* `chat:write.public` is to send messages to public channels without joining them first, and

* `channels:read` is for listing channels (and getting their ids).



This demo expects all the above permissions set (i.e. part of the Bot Token Scopes). We use

[slack](https://github.com/smallwins/slack) npm package, see its `README.md` file for documentation

and the list of available API along with the token scope(s) you'll need to set in order to

make use of given API call.



After setting the permissions, you can install the new application into your Slack workspace.

Finally, navigate again to Features / OAuth & Permissions and you'll get your access token there.



Add the token into `config.json` file under `slackToken` key, and configure the target *public*

Slack channel (its name) under `slackChannel` key.



### Slack integration: chat-bot



[Teaser.](http://recordit.co/iwBY1krysd)



On top of the webhook integration described above, you might want to add support for slash commands

that (like webhook notifications) also trigger data exchange with IP Fabric API server and follow-up

communication of the results back to Slack.



To activate slash commands, navigate to the settings of you Slack application, and in Features / Slash

commands, create new one. You will need to provide *public* IP address where the HTTP server listens

for the commands sent by Slack. The server in question is this agent, so you need to make sure it is

available on a *public* IP address.



Very likely, you have this agent installed near the IPF API server (or even on the same box), which

is typically inside a private network. To make it available on a *public* IP address, you might

use port forwarding with SSH, or need to configure your nginx forwarding.



### IP Fabric integration: API tokens



Sometimes the agent fetches additional data from IP Fabric API server (details about the reported

webhook event, data needed to process slash command from Slack). For this purpose, the platform

supports API tokens, and this demo also makes use of this feature. To make it work, please create

an API token in IPF UI, and store it under `ipfToken` key. To know what API server to talk to,

please specify the URL of IPF API server under `ipfApi` key in `config.json`.



### Development



The webhook agent is written using `koa` framework (similar to `express`, just better :-)),

and there is no tooling around it needed, as it is written in plain JS.



`src/agent.js` code is responsible for handling the webhook payload receiving part,

while `src/handler.js` code is the "what to do with the webhook payload" part.



### License



MIT