agent

README

Agent is the software that collects and deliver performance details to the Pinpoint Cloud

Building

You can build like this:

go install -tags dev

NOTE: Include the dev build tag do have access to the developer tools.

This will build the agent binary and install it in your $GOPATH/bin folder. Make sure this folder is in your $PATH.

Make sure and setup your $GOPRIVATE env to read any pinpoint modules from our repo:

go env -w GOPRIVATE=github.com/pinpt

Creating an integration

You will need to call the generate command and answer a few questions, for example:

> agent generate
    ____  _                   _       __ 
   / __ \(_)___  ____  ____  (_)___  / /_
  / /_/ / / __ \/ __ \/ __ \/ / __ \/ __/
 / ____/ / / / / /_/ / /_/ / / / / / /_  
/_/   /_/_/ /_/ .___/\____/_/_/ /_/\__/  
             /_/                         
Welcome to the Pinpoint Integration generator!

? Go Package Name: pinpt/github
? Name of the integration: Github
? Your company's name: Pinpoint Software Inc.
? Your company's url: http://pinpoint.com
? Your company's short, unique identifier: pinpt
? Your company's avatar url: https://avatars0.githubusercontent.com/u/24400526?s=200&v=4
? Choose integration capabilities: Issue Tracking, Source Code

🎉 project created! open ~/Documents/go/src/pinpt/github in your editor and start coding!

The project will be created. Go to that directory and open it in VSCode to start working on it. Happy coding.

Running for local dev

Clone the repo for the GitHub integration:

git clone git@github.com:pinpt/agent

Export

This will print each exported model to the console.

cd pinpt/agent
agent dev . --log-level debug --console-out --set apikey_auth='{"apikey": "$GITHUB_TOKEN" }'

You can run and have exports go to a directory if console-out is omitted. The default directory is ./dev-dir but you can pass in a --dir arg instead:

cd pinpt/agent
agent dev . --log-level debug --set apikey_auth='{"apikey": "$GITHUB_TOKEN" }' --dir exports

The --dir takes a folder to place the exported models (all data per model goes into one file JSON new line delimited).

If you are going to subscribe to webhooks, pass in the --webhook arg to register the url with out server, otherwise the register call will fail

cd pinpt/agent
agent dev . --log-level debug --set apikey_auth='{"apikey": "$GITHUB_TOKEN" }' --dir exports --webhook

Webhooks

cd pinpt/agent
agent dev webhook . --log-level debug --set apikey_auth='{"apikey": "$GITHUB_TOKEN" }' --input webhook_payload.json

The webhook_payload.json contains a sample webhook payload

Running for server

The server mode can either run in standalone or multi agent mode. Standalone mode is currently how agent's work today. Each customer has one instance of an agent. Multi agent mode is where the agent can act in a multi-tenant fashion and can process requests for multiple customers and these agents can be horizontally scaled.

Requirements

You must first build integrations:

cd pinpt/agent
agent build .

This will be placed in your dist folder as a file named github.

Standalone

Once you build the integration, you can just run it:

github --log-level debug --config agent.json

Currently, the agent.config format matches the current (legacy) agent config.

Multi

github --log-level debug

By default in this mode, will only talk with the local dev event-api. You can set --channel and --secret to point at another environment.

To place an event, use the event-api produce command such as:

go run . produce --log-level debug --channel dev agent.ExportRequest --input '{"customer_id":"1234","integrations":[{"name":"github","authorization":{"api_key":"XYZ"}}]}' --secret 'fa0s8f09a8sd09f8iasdlkfjalsfm,.m,xf' --header integration=github

Make sure you update the api_key with the value of your PP_GITHUB_TOKEN. Also, make sure you're running event-api server locally such as:

PP_CUSTOMER_ID=1234 PP_INTERNAL=1 make local

Record & Playback

You can enable record or playback of all HTTP interactions by passing in the --record or --replay arguments to dev. For record, this should be a directory to place the recording file. For replay, this will be the directory where the recording file was saved.

Contributions

We ♥️ open source and would love to see your contributions (documentation, questions, pull requests, isssue, etc). Please open an Issue or PullRequest! If you have any questions or issues, please do not hesitate to let us know.

License

This code is open source and licensed under the terms of the MIT License. Copyright © 2020 by Pinpoint Software, Inc.