Comparison
Data freshnessvsFreshness SLA
Data freshness
the dashboard renders perfectly and every number on it is from Thursday.
How old the data behind an output is, measured from the source event rather than from the last successful run. It is the metric that catches the failure mode nothing else catches: a pipeline that succeeds while producing nothing. Measuring it at the point of consumption rather than at the pipeline is what makes it honest, because every intermediate success is compatible with a stale answer.
Full entry →Freshness SLA
the agreement says this table is never more than three hours behind, so a four-hour delay is a breach rather than an opinion.
A stated commitment about how current a dataset will be, agreed with the people who depend on it. Its value is that it converts an unbounded expectation into a number that can be measured, alerted on and occasionally renegotiated. Without one, every delay is a conversation from first principles, and the pipeline team ends up defending latency numbers nobody ever agreed to.
Full entry →