This repository holds two things:
data/transition-sites- Configuration for sites being redirected to GOV.UK. This configuration is automatically loaded into the Transition app
tools/- Miscellaneous scripts for updating tna_timestamps, configuring root domain redirects and generating mappings.
Adding a new site
Before you start, you need to know:
- the domain name being transitioned and any aliases
- the organisation that owns the site (and any additional organisations that should have access)
- the new homepage for the old site - often this is the organisation's page on GOV.UK, but sometimes it can be a different page on GOV.UK
Create the site yaml file
Substitute each argument and then run the following rake task (you may need to run
bundle install first):
abbr should be of the form:
Example for the "Obesity West Midlands" site which is owned by Public Health England:
abbr is used in the URL for the site in the transition app. It is used as the primary key for a site and so shouldn't be changed once imported into the transition app, or a duplicate site will be created.
whitehall_slug is the GOV.UK slug of the organisation that owns the site. This determines several things:
- where the site is found in the Transition app
- who has access to edit and create mappings for the site
- what organisation name and branding is used on pages served by Bouncer for URLs which aren't redirected
Extra organisations can be added later.
host will be one of the hostnames that the site has, for example:
Aside: extra hostnames can be added later.
Adding extra config to the site
The generated file only contains fields which are absolutely required.
Check that the
tna_timestamp field is set. The rake task will try to find the latest archive by scraping the National Archives site, but this doesn't always work. You can find a list of the timestamps the National Archives have by going to their site, for example: here are the timestamps for this Cabinet Office site
Some of the optional fields are frequently used:
aliases- Almost every site will require aliases to cover the host name with/without
www.(depending on which host was used in the
options- used to specify when a site's URLs have significant querystring parameters
extra_organisation_slugs- used to enable members of other organisations to edit the mappings
See the Site configuration for more details about these and other fields.
Run the validation checks
These perform some cursory validations of the site configuration files.
bundle exec rake validate:all
An option for a more thorough test of the config is to import it into the Transition app in development.
Open a Pull Request
Commit the changes, push the branch and open a Pull Request.
Import the site to Transition
Once merged, wait for the site to be imported into the Transition app. This currently happens automatically on the hour via a job on deploy.production and deploy.integration Jenkins' during working hours. It can also be triggered manually.
Each site is configured using a yaml file in the
whitehall_slug— the slug used in Whitehall for the organisation which owns the site. Used for branding in Bouncer and access control in Transition.
site— friendly name for the site
host— primary hostname for site
tna_timestamp— timestamp of the last good National Archives capture. eg 20131002172858
homepage— URL for new site, used to redirect '/'. Must include the 'http' or 'https'.
homepage_title— site title for 404/410 pages. Defaults to organisation title. Should fit into the sentence: "Visit the new [title] site at [furl or homepage]"
extra_organisation_slugs— additional organisations which own this site. Used for access control in Transition.
homepage_furl— friendly URL displayed on 404/410 pages. Should redirect to the
homepage. Doesn't need to include 'http' or 'https'.
aliases— list of alias domains
global— set a global redirect or archive for all paths
css— a css class which determines the logo and brand colour used on 404/410 pages
options— used to list significant querystrings for canonicalisation like this:
--query-string first:second:third. A significant querystring parameter is one which on the old website changes the content in a meaningful way - which we might therefore need to map to a different place. Query string parameters should be specified in lowercase; uppercase parameters will not be preserved during canonicalisation.
global_redirect_append_path— should the path the user supplied be appended to the URL for the global redirect?
special_redirect_strategy— when the transition is partial, some tools or content will be left behind and managed by the previous supplier. This setting can be one of:
via_aka- the supplier is redirecting some paths to our aka domain.
supplier- the supplier is managing redirects to gov.uk. No traffic comes through Bouncer for this site.
We continue to serve some pages and assets for Directgov and BusinessLink sites. These are stored in GitHub and served by Bouncer's nginx configuration. See:
Bouncer's nginx configuration also includes a small number of redirects and other behaviours not possible with mappings.
Glossary of terms
A glossary of the terms used can be found in this blog post.