Reliability
Observability
Observability is the ability to tell what a system is doing and why from the data it emits, mainly logs, metrics and traces, so problems are found and diagnosed quickly.
How it works in practice
Monitoring answers whether something is wrong. Observability helps answer what and why. For automations and AI agents that means recording each run with its inputs, outputs, duration and errors, counting successes and failures over time, and for agents, recording each model call and tool call with its cost.
The common failure in small companies is silence: a workflow stops after a password change and nobody notices for weeks. We send failures to a named person in a channel they read, alert on volume dropping to zero as well as on errors, and keep run history long enough to answer questions from finance or auditors.
A worked example
An invoice sync normally posts 40 to 60 invoices a day. An alert fires if fewer than 10 post by 3pm on a weekday, which caught a broken connection on the day it happened instead of at month end.
Examples in this glossary describe typical builds for companies of 1 to 1,000 people. They are not named client results.















