The Navigator's First True Wind: On the Folly of the Fixed Bearing and the Living Tack
We build our services to sail the digital seas, and in our quest for reliability, we often install lighthouses. These are our uptime checks—steady, predictable pulses sent to a known endpoint, waiting for the expected ‘200 OK’ to flash back. They are our fixed bearings, a straight line drawn on a chart from A to B. They tell us if the rock is still there and if the light is lit. But they tell us nothing of the wind, the current, or the hidden reef just beneath the surface.
This is the folly of the fixed bearing. We become so focused on the endpoint’s mere existence that we forget the journey our users must take to reach it. A service can be ‘up’ by the strictest definition—responding to a simple health check from a monitoring node—while being utterly unusable. Latency can spike into the seconds for a user halfway across the globe; a critical API endpoint might return errors under load while the root URL hums along happily; a third-party service we depend on might slow to a crawl, poisoning our own response times. The fixed bearing sees none of this. It only confirms the rock hasn't moved.
The contrasting approach is what I call the living tack. It is the navigator’s art of using the wind itself, of constantly adjusting to the conditions of the sea. In our world, this is synthetic transaction monitoring. It is not a simple pulse but a scripted journey that mimics a real user’s path: logging in, adding an item to a cart, fetching a personalized feed, completing a checkout.
Where the fixed bearing is a binary question, the living tack provides a continuous stream of truth. It measures not just availability, but performance, functionality, and correctness at every stage. It discovers that while the login page is fast, the search API is timing out. It finds that a CSS change broke the checkout button on mobile viewports. It feels the sluggishness of a database read replica long before it becomes a full-blown outage. The living tack doesn’t just tell you the lighthouse is working; it tells you if your ship can actually sail to it.
Relying solely on the fixed bearing is a seductive trap. It offers the clean simplicity of a green/red status, a false comfort that encourages us to watch the monitor instead of the service. The living tack is inherently more complex, requiring care and maintenance to keep the synthetic journey aligned with real user behavior. But this complexity is its strength. It is an active, not a passive, form of vigilance. It acknowledges that a service is not a static point on a map but a dynamic, living system. True reliability isn’t just about being up; it’s about being truly, functionally, and performantly reachable. Don't just plot a bearing. Learn to tack.
Notes & further reading
A few pages I came back to while writing this: