GitHub

Connecting a repository is the alternative to reporting deploys from CI. RankDebug reads the repository's production deployments and records each one as a deploy, so drops can be lined up against releases without touching your pipeline.

Access

The connection installs the RankDebug GitHub app on your account or organisation. The access token it uses asks for three read permissions and nothing else:

• Contents: read
• Deployments: read
• Metadata: read

Connecting

1. On Connections, click Connect GitHub. GitHub asks where to install the app and which repositories it may see.
2. Back in RankDebug, choose the repository that deploys this website.

What is recorded

RankDebug looks for deployments to a production environment, meaning an environment marked as production or whose name starts with "prod", that finished successfully. Each one becomes a deploy at the time it went live.

If the repository has no production deployments, each commit on the default branch that touched SEO-relevant files stands in for a release instead.

For every deploy, RankDebug reads the changed files to find SEO-relevant changes: robots.txt rules, sitemaps, redirects, and lines that add or remove things such as noindex, canonical, hreflang, titles, meta descriptions and structured data. The diff is read and dropped. Only file paths and signal names are kept, such as "canonical removed, line 40". No line of your code is stored.

The repository syncs every hour and the first pull reaches back 90 days. A deploy older than six hours when it is found is recorded for history, but the site is not checked for it, since the site has moved on since then.