Croft Adds Built-In Monitoring and AI-Assisted Fixes to Every App

Summary

Croft says every app on its runtime now includes monitoring without an SDK, separate agent, dashboard setup, or additional cost. The service tracks response-time breakdowns, throughput, failure rates, grouped errors, slow requests and SQL queries, seven days of searchable logs, health checks, resource usage, and deployment markers. Its MCP interface exposes app errors, metrics, and logs so an AI assistant can inspect stack traces, identify failed requests, modify the code, and deploy a fix. Users can also select “Fix with Build” to send an error’s logs and stack trace to Croft Build, the company’s built-in AI chat. Error groups automatically resolve after a deployment produces no new occurrences. Croft sends email alerts for new errors, high failure rates, and unhealthy apps, with a limit of one alert per app per day and recovery notices for health failures. Logs and detailed traces remain on the customer’s Croft server for seven days; only aggregate counts and timings are sent to the control plane, while request bodies, query strings, cookies, and IP addresses are excluded. Monitoring is included in Croft’s Solo and Team plans without usage-based pricing or a separate bill.