I don’t use Sentry. I don’t use PostHog. I don’t use any dedicated error-monitoring SaaS.
Every single error in my stack - backend, frontend, queues, Durable Objects, etc, lands in Cloudflare Observability the moment it happens.
Then a tiny custom toolbar app pings my account every 15 minutes, surfaces any new errors, and I fix them typically within a few hours.
Over time the apps just… stop breaking.
Here’s the exact system (and why Cloudflare is pretty cool);