Weekly Netlify build health report across every site

Every Monday at 9am, build me a weekly deploy health report covering every site in my Netlify team, and email it to our engineering lead. This is a retrospective rollup across the whole team account, not an alert about a single broken deploy.

Start with Netlify List Team Sites to enumerate every site in the team. Do not report on just the site that failed most recently. For each site returned, call Netlify List Site Deploys and Netlify List Site Builds and gather the activity from the last seven days. Deploys come back newest first, so you can stop paging through a site's history once you reach records older than the window.

From those records, work out per site: the number of deploys in the window, how many succeeded versus failed, the resulting success and failure rate, and the average build duration. Derive all of this from the state field and the timestamps on the deploy and build records. Important: do not attempt to report build minutes consumed or bandwidth usage. Those come from Netlify's paid analytics feature, which is not available here and will simply fail for accounts without that plan. Deploy state and timestamps are enough for counts, rates, and durations.

Then reason about the numbers rather than dumping them. I want you to identify which sites got flakier this week compared with their earlier record in the window, which failures repeated across multiple deploys and therefore look like a persistent problem rather than a one off, and which sites have not deployed at all. Where several failures on a site share a pattern, say so explicitly instead of listing each failure separately.

Write the result as a readable report, not a table dump. Open with the two or three things actually worth acting on this week, stated plainly, before any per site detail. Then give the per site breakdown with the counts, rates, and average build duration. Finish with a separate section listing sites that had zero deploys in the window, called out as possibly stale or abandoned so someone can decide whether they are dormant on purpose. Send this using Gmail Send a Message to the engineering lead, with a subject line that includes the week ending date.

After sending the email, append one summary row per site to a Google Sheets spreadsheet using Append Values, so we accumulate a running history of build health over time. Each row should carry the week ending date, the site name, the deploy count, the success count, the failure count, the success rate, and the average build duration, in a consistent column order every week so the history stays comparable.