electricflow-examples

Different examples of how to use features in ElectricFlow.

Setting up examples

The requirement is that you are able to access the CLI and have signed on with ectool with permissions to change the system.

logon with "ectool login yourusername yourpassword"
You have administrative rights and your account lets you create projects, resources, and other objects.
This is not a production system and you are working on a system where you can make changes and mistakes.
You are able to create files on the local file system.

Keep these guidelines in mind:

Each example should reside in a separate folder under /examples, named something appropriately. As an example, the first example is named "SingleApplication" to convey it is about a single application. I suppose it could be called something like "SingleApplicationDeploy."
The example folder should contain a top-level "install.sh" script that installs the example.
- Let's assume plain old "./instal.sh" does a brand-new installation of the application.
- Re-invoking "./install.sh" should not produce an error.
- An dependent files should reside in this folder only.
- If you download from external sources, download them to the root-level (i.e. /vagrant) folder named downloads. Assume you can delete the contents of this folder and restart.
Document your application process in a suitable README.md at the top-level

The goals are:

Simplicity.
- Each item or widget should be managed managed separately, and not in a large uber-setup script. This means that if you have an example to showcase rollback, it can (and probably should) be separate from other examples.
- Each item or widget follows the same setup pattern - a self-contained directory with a setup script. Nothing too fancy and not too many levels of indirection.
Focused Examples
- Create standalone examples to show simple ideas. A deploy, a deploy with parallel steps, notification, rollback, rolling deploy should all be separate items.
- Think of this as an environment to answer "Ask" questions. For example, "How do I create a notification on a pipeline stage?"
- Also think of this as a way to showcase individual plugin eamples. For example, if you test a plugin integration for Amazon, create a subdirectory here to showcase the problem.
- In demonstrations, consider these contributions to be side-examples. "Here is an example of a rolling deploy using a simple application." The example is about rolling deploy and not about Tomcat, a plugin, notifications, or something else in particular.

Consider separating the projects to a public repository.
For each project, consider writing up a blog to describe how it works.
Need to figure out how to specify dependencies. For now, dependencies are just not handled. One example - using the same artifacts.
Where possible, we should be open to running with contemporary patches and update (i.e. apt-get update)
Figure out if it is worth including arguments to support clean, reinstall, install.
Consider adding support for each installation to specify project, or where the example will be installed. This will be a paramter to install.sh.

Add a notification to a failed deployment via DSL.
Add a notification to a pipeline stage via DSL.
Show how we can trigger an event from a GitHub checking (github.com)
Show how we can spin-up a cloud-based instance from EF
- Consider Azure, Google Cloud, AWS as examples
- Use the instance as an endpoint
- Or, use the instance as an agent