Configurable schedules Planned
Run a scenario every minute, every five minutes, or hourly. Different cadences for different criticality.
Monitoring · Coming soon
Continuous checks. Same forensic detail.
Run your scenarios on a schedule from your chosen regions. When LCP regresses or a step breaks, the alert comes with the session video, the network log, and the console output for the run that triggered it. You debug the incident, not just react to it.
Monitoring is on the roadmap behind Testing Suite. Same scenarios, run continuously.
Monitor schedule mockup
1380 × 1035 · mockup
What it does
Synthetic monitoring, with the report attached.
Most synthetic monitoring tools tell you a page took 3.2 seconds to load. Then you go to find out why. Open a different tool. Look at RUM samples. Cross-reference with deploy times. Eventually find that a third-party tag started timing out.
Monitoring in Evaluat runs your scenario in a real browser every N minutes, from a chosen region. When the run fails, or LCP regresses past your threshold, the alert comes with the session video, the network log, the console log, and the step playback for the run that triggered it. You debug the incident from one page.
Same scenarios you use for performance tests. Same configuration UI. Same report. Different schedule and concurrency. Build once.
What an alert looks like
The alert IS the debug session.
Slack message: "LCP on /product/SKU-12345 regressed from 2.1s to 3.8s on the Frankfurt monitor at 14:22 UTC. View session."
Click through. You're in the session report for the run that triggered the alert. Video. Network. Console. Step playback. The third-party tag, the slow database, or the broken deploy. It's already there.
Roadmap
What's planned.
Rough shape of what Monitoring will ship with. Subscribers get an email when each piece is usable.
Multi-region runs Planned
Run the same scenario from London, Frankfurt, and the regions we add next. Compare what users see in different geographies.
Threshold alerts Planned
Alert when LCP, INP, CLS or a custom step duration crosses your budget. With the session attached so you can debug.
Slack & PagerDuty Planned
Alerts that land in the right channel, with links to the failing session video and step playback.
Trend dashboards Planned
Long-term Web Vitals trends per scenario and per region. Spot the slow regression that aggregate RUM misses.
Public status feed Planned
Optional public status page powered by your monitors. Customers see real performance, not "all systems operational."
What teams use it for
Six monitoring patterns we hear on demo calls.
Reasons to point a continuous real-browser monitor at your critical journeys. Every alert ships with the session.
Regional baselining
Run the same checkout monitor from London, Frankfurt, São Paulo. Catch the latency that only your German customers see.
Slow regressions
Aggregate RUM hides the 200ms LCP creep across a month. A continuous synthetic monitor shows it. With the session video for each data point.
Step-level failure alerts
When the login step starts failing because a CDN-cached asset went 404, the alert names the step and links the session. No more "site is down" pages.
Third-party watchdog
Monitor the third-party tags you embed. Alert when the analytics SDK adds 800ms to TBT. Catch it before the customer support tickets do.
Public status fact-check
Don't say "all systems operational" when a customer can't complete checkout. Power your status page with monitor results, not vibes.
Deploy-correlated alerting
When a Web Vital crosses the threshold within minutes of a deploy, the alert links to the regression candidate. Cause and effect, attached.