Operations dashboard: a single page that tells us if we’re healthy
Running a small lender well is partly about reading the dashboards before customers tell you something is wrong. This week we shipped a new internal page — accessible only to the company director — that condenses every dependency the credicorp.co.uk site relies on into a single colour-coded checklist. We do not publish the page externally, but the principles behind it are worth explaining, because they reflect how we think about availability and trust.
What it tells us
The page is a simple table of green / amber / red lights, one row per dependency. PHP and WordPress versions, the database connection, the outbound mail service, our Anthropic API key for the customer assistant, our VAPID configuration for push notifications, the decision engine’s last-fired timestamp, maintenance mode state, and whether the fatal-error log file has been touched recently. Below the dependencies is a small block of counters: number of customers, applications and loans, number of confirmed newsletter subscribers, number of active push-notification subscriptions, and how many of our staff have been active in the last half hour.
Why a single page matters
Outages are rarely caused by the thing you were watching. They are caused by the thing you had not checked in a while. A page that consolidates everything onto one screen, with no need to open six tabs, is a small change that lets us catch most surprises before customers see them. It does not replace proper alerting, but it makes the morning round much faster.
What it does not do
The dashboard is read-only. It probes services on demand (so a check that is amber today may be green five minutes from now once we refresh), and it does not include anything customers would consider personal data. It is an operational health view, not a business-intelligence view.
The wider posture
We try to write down both our policies and the small operational habits that keep them honest. The dashboard is a habit. Other examples include the audit log that records every staff impersonation event and the always-on banner that shows when we have demo mode enabled. These small surfaces stop us drifting from our own standards over time. The director reviews the dashboard every morning and during any change window; if you ever wonder what level of attention we apply to keeping the service healthy, this is part of the answer.
