問題文
A platform team wants to know whether their Collector fleet is losing data before it reaches the ingest endpoint. They have not built anything for this yet. Which approach gives them that answer with the least custom work, and what specifically should they look at?
選択肢
- Write a custom receiver that duplicates every data point and sends one copy to a second organization, then compare the two organizations to find the difference, because there is no other way to observe loss at the Collector itself.
- Enable a higher log level on every host and read the logs by hand, counting the send attempts and the acknowledgments for each Collector over a fixed window and comparing the two totals.
- Compare the number of hosts shown in the navigator against the number of hosts in the inventory spreadsheet each morning, because a host that is dropping data disappears from the navigator within a few minutes.
- Open the built-in dashboard for the Collector, which is fed by the Collector's own internal metrics, and treat any non-zero value in the section about dropped spans, data points, and log records as something to investigate.