Comparisons

Two tools, two questions. Neither column wins — what decides is which question you actually have in front of you.

Metrics vs Traces

Metrics tell you *that* something is wrong across all requests; traces tell you *where* the time went in one request. Neither substitutes for the other, and teams that buy only one spend years compensating for the gap.

Best for

Trends, rates, alerting and "is this normal?" across every request

Blind to

Why any individual request was slow, and anything not pre-aggregated

Costs

Cheap per request; cost explodes with label cardinality

Answers

"Are we slower than yesterday, and for which endpoint?"

Best for

Locating latency across service boundaries in a specific slow request

Blind to

Aggregate behaviour, unless you sample carefully and reason statistically

Costs

Expensive per request; usually sampled, which risks missing the tail

Answers

"Which span consumed 1.8 of these 2.4 seconds?"

DimensionMetricsTraces
GranularityAggregated across all requestsOne request, end to end
Cost driverCardinality of labelsVolume and sampling rate
Finds the tailShows that a tail existsShows what is in it — if sampling kept it
AlertingThe natural fitRarely alerted on directly
Cross-servicePer-service, joined by handJoined by construction