Highlights
Dashboard first-run looks completely different. If you just signed up and haven't sent your first request, you'll see sample data, a working curl snippet, and empty states that actually tell you what to do next. We also rebuilt the status page and made list endpoints respond noticeably faster across the API.
What's New
Dashboard: first-run experience overhauled
Before this week, signing up and landing on the Overview was awkward. Empty charts, empty tables, no clear first step. We fixed that.
New API keys now show an onboarding empty state with the exact curl you need for your first request, including your real key and the right endpoint. The Overview shows ghost-mode sample data so you can see what the dashboard looks like in action before your first call lands. Banner copy is in English everywhere (we caught a few stray mixed-language bits).
Small changes. But first impressions matter, and you shouldn't have to guess what to do after you sign up.
Status page: rebuilt so it reflects reality
We rewrote our status page this week so it stays accurate after restarts and across deploys. No more stale events lingering, no more uptime numbers padded by missing data.
What you see on the public status page:
- Daily bars render grey on days where we don't have monitoring data, instead of green (we were counting blind days as success; we're not anymore)
- Hover tooltip shows minutes per severity, so you can see exactly how much degradation happened
- Uptime percentage excludes blind days, so the number actually reflects reality
- Customer-facing incident titles read in plain language, not engineering shorthand
The public status feed now reflects live monitoring data, the same view our team watches, sanitized before it lands on your screen.
Faster list endpoints
List views across the API now respond noticeably faster, with the biggest gains on Proxy Finder.
We made our backend treat list views differently from detail views. List endpoints return narrow records by default, and the heavy fields stay in detail responses where you actually ask for them. We also cleared a stale sort hint that was forcing one of our common queries to scan more data than it needed.
Net effect: less waiting on long lists, the same data when you fetch the full record.
Rate-limit outcomes now show up in analytics
If a request hit an upstream rate limit, that outcome wasn't showing up in your per-domain and per-IP analytics. Fixed. Rate-limit is now a first-class outcome you can see and filter on, alongside success, failure, and timeout.
If you've ever wondered why the numbers didn't add up on heavily protected sites, this was part of the reason.
Under the Hood
Analytics on the dashboard should feel more reliable now. We were hitting occasional resource ceilings during peak traffic that turned into flaky chart loads on the Overview. That's behind us.
We also finished separating the events our team watches behind the scenes from the incidents that belong on the public status page. You only see the second one, written in plain language.
If a specific change here caught your eye and you want the deep version, keep an eye on the Update category. That's where the longer write-ups land.