#circle ci

[clye]

[ox86]

CI as a service: More funky stuff CircleCI can do

In my recent post about CircleCI I pointed out how easy it is to get started with Continuous Testing thanks to CircleCI.

After a few mails to and from the customer support (Thanks to Paul Biggar and Allen Rohner!) I found out, you can do a lot more, than I suspected:

Deployment

A vital part of fully leveraging Continuous Integration for me is being abled to automatically deploy (either to a live or a staging environment) what has been identified as a working build - or Continuous Deployment in shorter terms.

After browsing through the project settings and looking around in the platform, I only found a way to deploy to Heroku automatically and concluded, that other deployment strategies are not supported for now.

This is a case of RTFM, I have been stupid here: You can deploy anyway you want, as the docs point out.

API

The curious child I sometimes am I wondered, if there is an API, like the Jenkins API, to have something on our wall displays, yelling at people if builds fail (as we have using the mentioned Jenkins API).

There are two ways of accessing the data from CircleCI:

A notification hook: This is a URL in your control where CircleCI sends information about the project and its build state to (as JSON). You can put some server-side script there and do something with the data then.

An initial version of a REST API! This one is undocumented and pretty limited for now - but promising!

Github Commit Status API

Github has published their "Commit Status API" which, to put it short, allows marking individual commits as "working" or "broken".

That Circle CI supports this is pretty fancy.