Why should label planning precede a large Prometheus metrics import?
Select an answer to reveal the explanation.
Short Explanation
You need to plan labels before a large Prometheus import because cardinality is the cost driver. Every distinct label combination becomes another stored time series, so a label like request ID can create one series per request. That inflates storage and slows queries.
Full Explanation
The correct choice is the option that says labels multiply the number of stored time series, so unplanned high-cardinality labels inflate storage and slow queries. Each unique combination of label values is a separate time series. Labels that carry unbounded values - user IDs, request IDs, full URLs - generate enormous cardinality, which drives storage growth and query cost.
The option claiming Instana rejects any metric that carries more than one label is incorrect because multiple labels are normal and expected; the concern is which values they can take.
The option claiming labels must exactly match Instana's internal entity names or ingestion fails is incorrect because labels do not have to mirror Instana's entity naming, though mapping them thoughtfully improves usability.
The option claiming labels are removed during ingestion, so planning them determines nothing is incorrect because labels are retained and are what makes the metrics filterable and groupable.
For the exam, connect label planning to cardinality and cardinality to storage and query cost.