問題文
Spans stop arriving entirely a few minutes after a deployment, while metrics from the same process continue. Which cause is consistent with that pattern?
選択肢
- The Resource being invalid.
- A failure of the whole SDK, because the provider initialization is shared between the signals and therefore a failure in one of them stops the others as well.
- The propagators variable being set to none, which stops the trace pipeline while leaving the metric pipeline alone.
- A signal-specific setting for traces, since exporters, endpoints and timeouts can all be configured per signal.