PCA (Prometheus Certified Associate) sample exam question with answer 250

Question:
Given the following time series and assuming a scrape interval of 1m, what is the result of the query given below?
Series:

up{job="node", instance="node1"} 0 1 1 1 0
up{job="node", instance="node2"} 1 1 1 1 1

Query: sum(up{job="node"}) / count(up{job="node"})

1. up{job="node", instance="node1"} 0.6
   up{job="node", instance="node2"} 1

2. {job="node"} 0.5

3. {job="node", instance="node1"} 0.6
   {job="node", instance="node2"} 1

4. {} 0.5

Answer:
D - is the correct answer
A - is incorrect as both the sum and count aggregations drop the metric name from the result
B - is incorrect. Though the value is correct, the labels are not. Since no labels were specified for the dimensional aggregations sum and count, they aggregate across all dimensions. Any labels for dimensions that are not explicitly aggregated across are dropped from the label set of the result. For this to be the correct answer, the query would have to be something like sum by (job) (up{job="node"}) / count by (job) (up{job="node"})
C - is incorrect as both the sum and count aggregations act over dimensions, not time. Thus, only one result series is produced. This would have been the correct answer if the query was avg_over_time(up{job="node"}[5m]