Skip to content

jeremy-lq/documentation

 
 

Repository files navigation

Documentation site for Datadog

Built with nanoc, a static website generation tool.

Setup

Install Rbenv and Ruby 2.3.0:

brew install rbenv # or equivalent on linux (if you don't know about brew, go to http://brew.sh)
rbenv install 2.3.0
rbenv local 2.3.0

To configure this Ruby to take precedence over the system Ruby, add the following to your ~/.bash_profile (or similar file for your shell):

eval "$(rbenv init -)"

Open a new shell or source your ~/.bash_profile and install bundler:

gem install bundler

One dependency of the project - the therubyracer gem - requires a specific of the V8 Javascript engine. Install the correct V8 first and then install version 0.12.2 of therubyracer:

brew tap homebrew/versions
brew install v8-315
brew link --force v8-315
gem install libv8 -- --with-system-v8
gem install therubyracer -v '0.12.2'

Then, from the root of the project repo, install all the gems:

rbenv exec bundle install

As of OS X 10.11 (El Capitan), OpenSSL headers are no longer provided. You will need to install OpenSSL prior to running the bundle install.

brew install openssl
brew link openssl --force

If you are running OS X 10.12 (Sierra), linking OpenSSL as detailed above will throw a warning and the bundle install will fail. To link the OpenSSL headers, run the following after setting the rbenv local version and prior to running the bundle install.

brew install openssl
rbenv exec bundle config --local build.eventmachine --with-opt-dir=/usr/local/opt/openssl

Integrations that have metrics will require your Github Personal Token. For more information on generating a token, see Github's documentation. After you've generated a token, add the following line to the .bash_profile in your home directory:

export github_personal_token=[paste access token here]

You should then run source ~/.bash_profile to reload the settings.

Working on Docs

rake

If you get an error, make sure you're running on the host.

Yeah, that's it. This command will compile the site, check for any bad links, and refresh your browser.

This site uses Kramdown. To learn about the syntax, see this site.

If you include ANY Markdown in a file, give it an .md extension.

Make sure all files are lowercase. Macs are case insensitive when creating links to images and pages, but the server is not. The tests will be fine locally but the site will fail in production.

Releasing

Before push/merging, make sure to

rake clean
rake

and verify that there are no bad links on http://localhost:3000.

If you've been working on code samples you should also

export TEST_DD_API_KEY=test_org_api_key
export TEST_DD_APP_KEY=test_org_app_key
rake clean
rake test

These keys should be for a test account that does not include dozens of people. There are several samples that mute and unmute everything. Everyone in the org will be notified. If you are the only one in the org, you won't be getting angry emails from others asking you to stop muting everything.

If you have a super slow connection, perhaps on the Amtrak or on a plane, use

rake slow

This command is the same as rake but will skip getting metrics from various repositories. Running rake on a train can take 500 seconds, but changing this to rake slow makes it less than a minute for a full compile.

If there are errors, please don't merge.

If you receive an error regarding There was a problem getting GitHub Metrics, please see the Github personal access token information in the setup instructions above.

Internal Datadog folks: Within 5 minutes of merging to master, it will deploy automatically. You can see the status in #documentation.

How to add a new integration

Integrations include information that comes from two different sources. The first and main source are the files under content/integrations. The second source are the metric csv files under dogweb. In order to see the metric tables that appear for some of the integrations, you need to create an environment variable called github_personal_token assigned your github personal token (you need to have access to dogweb and therefore must be a Datadog employee to see this). The table will be brought in automatically by the deploy process.

The top of each integration file should include the following frontmatter:

---
title: Datadog-<integration name> Integration
integration_title: <integration name>
kind: integration
doclevel: basic
---

If you are writing a lot about the integration, change doclevel to complete or just remove the whole line. Now write the doc. There is no need to update any index, menu, or sidebars. Those are automatically generated.

Most integrations start with a heading level of 3. Going forward you should start with 1. But if you do, make sure to add the newhlevel attribute to the frontmatter: newhlevel: true

Every integration should have the following format:

Overview

Absolutely Required.

The first thing in the Overview should be a representative image for the integration. Try to make it as interesting as possible.

The overview section is required and should be a paragraph or two with some bullets of what is interesting about this integration. For example, the following comes from the Docker integration.

Get metrics from Docker in real time to:

  • Visualize your containers' performance.
  • Correlate the performance of containers with the applications running inside.

There are three ways to setup the Docker integration: install the agent on the host, on a single priviledged container, and on each individual container.

Installation

Required with some exceptions

The installation section should cover anything that needs to be installed on the agent host. For instance, in the Docker installation section you learn about installing the agent into a container. If there is nothing to install on the agent host, this section can be left out. To be a complete integration, either an installation section or a configuration section must be included.

Configuration

Required with some exceptions

The configuration section should cover anything that you can configure in the Datadog interface or the agent configuration files. In almost every case this section should be included since there is almost always something to configure. To be a complete integration, either an installation section or a configuration section must be included.

At the end of the configuration section include a link to the example configuration files. This should be done by adding <%= insert_example_links%>. This method takes a few optional parameters: conf is the name of the example YAML file, minus the extension; check is the name of the check file, minus the .py extension; setting either check or conf to "none" will hide that line; include_intro set to false will show only the list minus the sentence at the top; normally the integration title in the links will come from the pages frontmatter, but setting integration will override that, yaml_extension will change the extension from example to something else (like "default").

Configuration Options

Describe each of the options available in the YAML file. This will often be the stuff included in the YAML comments (remove them from the YAML included in the doc), but sometimes you will have to investigate a bit to figure out what the option is for.

Validation

Required

The validation section should include instructions on how to validate that the integration is successfully working.

Metrics

Required for integrations that have metrics

If the metrics are listed in the integration under dogweb, add an attribute to the frontmatter: git_integration_title: integration_name replacing the integration name with the name of the folder for the integration in the dogweb repo.

Then add <%= get_metrics_from_git()%> to the Metrics section. This will use your Github Personal Token to grab the metrics from the repo. For more information about setting up your Github Personal Token, see the Setup section above.

Events

Optional

This is a newer section. If you can describe the events that are part of the integration, it should go here.

Troubleshooting

Optional

The troubleshooting section should include anything that answers a question a user might have about the integration. If there is a question that comes up in support about the integration, it should be added here.

How to add a new Guide

Create a markdown file under content/guides. Add the following front matter at the top of the file:

---
title: <guide title>
kind: guide
listorder: <where in the list you want the doc to appear>
---

Each guide has a listorder. Change the list order number of this doc and any other docs to make sure stuff appears in the right order. There is no need to update any index, menu, or sidebars. Those are automatically generated.

About

The source for Datadog's documentation site.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • HTML 57.1%
  • Ruby 16.3%
  • Shell 9.8%
  • Python 8.1%
  • JavaScript 5.2%
  • CSS 3.3%
  • Other 0.2%