Roundup: JavaScript sites that read their own logs

Roundup: JavaScript sites that read their own logs

Three case studies on what Googlebot actually fetches on a single-page app.

→ Onely render-budget study — logs showed Googlebot requesting the JS bundles only 8% as often as the HTML shell. The gap explained why fresh content sat unindexed for 11 days. Server-rendering the key route cut time-to-index to under 2 days.
Takeaway: if the bot isn't fetching your JS, it isn't seeing your content.

⭐ Pick of the week — Merj / Google logs talk — separated Googlebot's HTML crawler from its rendering fetches in the log, showing they run on different schedules. Numbers: rendering lagged HTML crawls by an average 9 days on their sample.
Takeaway: 'crawled' and 'rendered' are two events, two log lines.

→ web.dev rendering guide — pairs nicely; explains why the second wave exists.
Takeaway: context for the logs above.