Lovable experienced two availability incidents on Friday, both of which only affected the main Lovable site. Published user applications hosted by Lovable were not impacted.
9:20-10:10 UTC: The Lovable website was slow to load, with some users seeing loading times up to 20 seconds. This was due to an outage at Posthog, a third-party vendor we use for running experiments and controlling feature roll-out. Since Lovable fetches configuration from Posthog before displaying content, the vendor's slowness caused a significant delay in the site's startup. See Posthog's incident report.
12:00-12:45 UTC: Our main website hosting vendor, Netlify, had an incident. This made the lovable.dev website completely inaccessible until Netlify restored their service, about 45 minutes later. See Netlify's incident report.
While both incidents stemmed from problems with upstream vendors, we can implement several mitigations to make Lovable more resilient to similar failures in the future. We were already planning to switch to local resolution of feature flags and evaluate contingency options for hosting, and as a direct result of Friday's incidents, we are now prioritizing this work.
