I have a situation where multiple ec2 instances has the same hostname.
So when i run promql on node_exporter metric with {instance=a} and if i have two instances with instance name a, i get only one result back. Not sure how the query is aggregating the result.
For a metric to be ingestible it has to have unique labels. So there must be a different set of labels for these two EC2 instances (otherwise you’d have errors about dropping duplicate metrics).