FLOW LGX · PROBLEM SOLVED · OCTOBER 2026
Once real companies depend on a platform, “is everything OK?” becomes a daily question. Server logs answer it for developers, not for the business owner. Flow LGX needed a page that answers it for both.
Month by month, how a logistics platform for a Toronto client went from a simple brief to a live, multi-company SaaS product.
A live platform needs a health check that anyone can read. Flow LGX has a monitoring page with a performance score, errors and slow requests, and the owner’s dashboard flags risks such as admin accounts without two-factor login.
Problems on a live platform rarely announce themselves. A slow search, a backlog of warehouse scans, or an admin account without two-factor login can all go unnoticed until a customer complains. The owner needed to see these signals without asking a developer to read logs.
The platform’s AI Ops page summarises health for any time window — the last 24 hours by default, or a chosen date range and event type. It shows a single performance score, the number of API errors, the average response time, the 95th-percentile (P95) response time, and queue or scan issues, including open scan discrepancies. Below the numbers, a list of individual events shows what happened and when.
Averages hide pain. If most requests take a fraction of a second but one in twenty takes ten seconds, the average still looks fine — yet the people hitting those slow requests are usually the ones doing the heaviest work, like loading a large jobs list. P95 means “95% of requests were faster than this”, so it shows what your slowest users really feel. When P95 rises, we know which screens to speed up first.
Health is not only speed. The super admin dashboard opens with an “alerts require attention” panel that lists risks such as admin accounts without two-factor login. Putting it on the first screen the owner sees each morning means it gets fixed, rather than sitting in a report nobody opens.
Give owners health signals in plain words on the screen they already open. A risk shown on the dashboard gets fixed; a risk in a log file waits for an incident.