Skip to main content
LogBrew uses one term for each product concept. API fields, CLI flags, enum values, event names, and attribute keys remain unchanged in every language.

Observability signals

A grouped problem that can contain multiple occurrences. An occurrence is one captured instance of that issue.
One timestamped event record. A log can carry severity, service, release, environment, trace context, and privacy-bounded attributes.
A trace follows work across a request or operation. A span is one timed operation inside that trace. A correlation to a trace is evidence of shared context; it is not proof of cause.
A runtime measurement with a defined unit, type, aggregation, and time window. Metric points are not interchangeable with log, trace, or issue counts.
A deployed application or service version reported with telemetry. A release is not the same object as a project or repository.
A classified product event, such as a signup, checkout, import, or failed job. An action is not an automated command that LogBrew runs.

Scope and data quality

Evidence states

These states are not interchangeable:
  • Captured: the evidence is present.
  • Missing: the evidence was not captured.
  • Redacted: the evidence was deliberately concealed.
  • Truncated: the evidence exceeded a response or capture bound.
  • Unavailable: the read could not provide the evidence.
Likely cause and likely fix area describe bounded diagnostic evidence, not a proven root cause. Reported impact is supplied by the reporting application unless a contract explicitly describes a LogBrew measurement.

Product analytics

Partial periods, missing identity, incomplete property capture, sampling, and truncation qualify the result. LogBrew surfaces those limits rather than turning them into zero or certainty.