Posted by maddhruvhn 5 hours ago
I am building SupaGuard. I am a software engineer and have been deeply involved in production observability for quite some time.
The Problem
I found out that Real User Monitoring (RUM) is the one that lets us down the most when we are in need of help. Taking an example, if a deployment at 7 PM breaks the login flow but no one tries to log in until 11 PM, then the system is operating blindly for more than 4 hours. Metrics might be showing that everything is fine, but in reality, the product is functionally broken.
The Solution
I am making SupaGuard to address only these kinds of pain points. It is an AI-native continuous synthetic monitoring platform that aims at locating those silent functional failures which do not generate any noise.
Key Features:
AI Test Generation: A browser agent goes through your app, automatically discovers critical flows and hence generates Playwright scripts along with assertions.
Smart Retries: To handle flakiness, when a test fails in a certain location (e.g. US East), it will try to rerun the test in another location (e.g. EU) immediately and only if the follow-up fails will it send the alert.
Non-Binary Alerting: It is not only "down" that is checked. The system grades performance degradation and identifies functional failures impact-wise.
Pricing
I am figuring out a better pricing model that will still be affordable and user-friendly at the same time. Suggestions are welcome.
Status
I am still in the process of product development and you can sign up for a waitlist if you'd like to be the first to know when it's ready.
The URL is https://supaguard.app/
I would be very happy to get your feedback regarding the concept and the features that I am prioritizing. Thanks!
Apart from this would you trust AI-generated browser tests to identify critical user flows in your app, or would you require manual curation also?
The AI Generated Browser tests are more like assistance for teams and people who don't want to have granular knowledge of Playwright scripting, users would be able to see recordings of tests and modify using natural language :)