Kickstart A New Project

This section is a sort of walkthrough on what have to be done - tools and data wise - to claim a project is "ready to start".

This process is in charge to the team leader, with the aid of accountant and sales.

Toggl

In charge to: Accountant

Information about contracts are no held by Redmine, but by Toggl instead.
This means team-members have to know which contract they have to track time onto.

To achieve this, we use the Project/Task feature of Toggl, following this schema:

Customer
  |
  -- Project
       |
       -- Task (Contract) 1
       -- Task (Contract) 2
       -- ...

For example, our the customer ACME wants us to build and maintain both their corporate site and their sales-supporting REST API.
We agree for four orders:

2016081101 - New corporate site development
2016081101 - Corporate site maintenance
2016081103 - Sales REST API development
2016081104 - Sales REST API maintenance

(please, mind the different offers ids)

the structure of project-tasks in Toggl may become

ACME
  |
  -- Corporate Site
  |    |
  |    -- 22016 development [016081101]
  |    -- Maintenance [2016081102]
  |
  -- Sales REST API
       |
       -- v1.0 Development [2016081103]
       -- Maintenance [2016081104]

The keys are:

The task partains to the right project for the right client
The task name contains a reference to the offer id
The task name contains an intuitive description for the developers

Redmine

Create main "Customer" project (if need be)

In charge to: someone who is Administrator on Redmine

Redmine allows projects to contain subprojects.
To keep things well organized we use to create an parent project for each customer so that:

Top-right select menu is easier to navigate for those of us who see a lot of projects
All information partaining to a customer, which impacts on all projects for that customer (say access credentials, contacts, workflow modifications, customer-related policies, etc) can be stored in the main project's wiki.

We generally don't want subprojects to have Wikis: project documentation is hold in the project's repository using our Raneto-based docker image - hint: if you start the project out of a base repository of ours, all is already in place.

If the customer is a new one, you'll have to create a main project for it. Here's how:

@todo insert screenshot of a project form properly filled

Name you project as the customer (say: ACME Ltd or Great Company Inc.)

Populate the Description textarea with this template

*Main project*
This holds all customer's related subproject and the general customer's Wiki

The Identifier fiels is important: choose a short machine-name for this project. Ideally it can be as short as the client initials, since it will prepend all subprojects identifiers. (i.e. ACME Ltd. → acme, Great Customer Inc. → gc or grcu)
The Public checkbox must be left off, as well as the Subproject of and Inherit member
In the Modules subsection, disable all modules but Wiki and Wiki extensions, unless your main project refers to an agency/provider (so different subprojects refer to different final customers and will have their own wiki)

Create project

In charge to: someone who is Administrator on Redmine

Create new project on Redmine, with the following settings:

Name you project in a meaningful way New site is bad, Corporate site redesign 2016 is good. Also avoid putting the customer's name in it again (say ACME - Corporate site redesign 2016), since the project is namespaced in its identifier (see below)
Populate the Description textarea as you see fit. Mind to not disclose information if the customer must have access to the project.
Populate Identifier field with a short and matching machine name, namespaced by the main project's identifier (in the previous example it was acme or grcu, so acme-corpsite-redesign-2016 or grcu-sales-rest-api are good examples)
The Public checkbox must be left off
Select the right customer's project in Subproject of select and check Inherit member as you see fit

Enable standard plugins

In the Modules subsection:

Make sure Wiki and Wiki extensions are disabled, unless your main project refers to an agency/provider (so different subprojects refer to different final customers)
Make sure Issue tracking, Time tracking and Agile plugins are enabled
Activate other plugins as you see fit (leave disabled in doubt, less fuss, less troubles)

Add members to the project

Once the project has been created, visit the project's Settings page (last tab in the project's menu), Members subtab.

Add all project members, with those roles:

Manager to Team Leader (he will make other senior dev managers if the need arise)
Developers to other team members

Create first project version

@todo move this introduction to a non-recipe file in the playbook!

We don't develop products so we actually don't need "versions" in the classic acception, like semantic versioning.

We then use versions to keep the pace of project development milestones. We have two policies depending on the management strategy.

Scrum and scrum-like projects use versions to mark Sprints
Kanban projects use versions to group issues that belong to the same requiremet group (may be a deploy date or a broad-lines requirements bucket)

No matter what, define a first version for your project:

Visit the project's Settings page (last tab in the project's menu), Versions subtab
Click on New version link at the bottom of the list (which so far should be empty)
Insert SPRINT-1 for scrum-like projects or an identifier for the first milestone/group (i.e. 2016-07-31 Deploy) for kanbans
(Optional) Set the end date, so we know when the sprint is supposed to finish or when the expected delivery is
Save and done

Slack

In charge to: Team leader

Open a Slack channel for the project in the form <client id>-<project id> (i.e. #acme-site, #grcu-api) and invite all team members.

In case of small projects, maintenance or if it sounds good enough, a simple channel with the customer's name is ok (i.e. `#acme) but please avoid having too noise in the same place!

In case external developers, customers, partners or other people have to access Slack, open a whole new slack with the name SparkFabrik - <project name> or SparkFabrik - <customer name> and invite all involved people.

Add project info helpers

Go to https://sparkfabrik.slack.com/customize/slackbot to add a response to info-gathering triggers
Add responses as appropriate (for example)

When someone says	Slackbot responds
help acme	available commands: envs, docs
help acme envs	Jenkins: http://ci.site.acme.sparkfabrik.com/ (night/3v0lv3_n0w) - Dev: http://dev.site.acme.sparkfabrik.com (stage/stage)

Add Gitlab integration

Go to app config on slack (ie https://sparkfabrik.slack.com/apps/A0F7XDUAZ-incoming-webhooks)
Click on "Add configuration"
Create a configuration and copy the "Webhook URL" field content
Go to gitlab setting for your project (example https://gitlab.sparkfabrik.com///services) and click on "Slack"
Click on active and paste the url into "Webhook" and save

Team group on GMail

In charge to: management representative

A group to broadcast mails to the team must be created on GMail.

Ask someone with the owner permissions on Sparfabrik's Google Apps for Work account to create a group with the following properties

The group name must follow the template: SparkFabrik - <Project> Team (mind the case!)
Add all Team Leaders and management representatives who are involved as onwers of the group
Add all developers as members of the groups

IMPORTANT NOTE: by convention this group contains SparkFabrik team members only. Do not add external users to the group to avoid leaking of classified information.

Last updated on 5 Sep 2024