I never realized how sloppy my use of observability terminology was until I read the OpenTelemetry documentation.
The docs offer concise and clear definitions, so I won’t duplicate them here. Check out the observability primer and the metric docs. They take only a few minutes to read and set a useful baseline for understanding observability.
I’d particularly like to highlight spans. Spans define units of work that group together a series of smaller operations and logs. Nested spans add up to a trace, which allows us to observe the lifetime of a request across process boundaries. It’s like having a stack track that works across service boundaries. Traces and spans greatly improve context of logs and improve diagnostics.