appengine.googleapis.com causes errors in prometheus handler

Hello,

When we include appengine.googleapis.com in our prefix list we see 500s being returned from the /metrics endpoint. We've turned debug logging on and it seems to be loading the app engine metrics without error:

time="2018-10-18T19:02:01Z" level=debug msg="Listing Google Stackdriver Monitoring metric descriptors starting with `appengine.googleapis.com`..." source="monitoring_collector.go:210"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/system/network/sent_bytes_count`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/http/server/quota_denial_count`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/system/memory/usage`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/system/network/received_bytes_count`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/http/server/response_count`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/http/server/dos_intercept_count`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/flex/disk/write_bytes_count`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/http/server/response_latencies`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/memcache/centi_mcu_count`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/memcache/operation_count`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/memcache/received_bytes_count`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/memcache/sent_bytes_count`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/system/cpu/usage`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/system/instance_count`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/flex/cpu/reserved_cores`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/flex/cpu/utilization`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/flex/disk/read_bytes_count`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/flex/network/sent_bytes_count`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/flex/network/received_bytes_count`..." source="monitoring_collector.go:166"
time="2018-10-18T19:02:01Z" level=debug msg="Retrieving Google Stackdriver Monitoring metrics for descriptor `appengine.googleapis.com/http/server/response_style_count`..." source="monitoring_collector.go:166"

There are no error logs coming from the prometheus handler (or any error logs coming in general) and I can't seem to find a way to turn on more detailed logging for the handler (new to Go so may be missing something).

Any idea what could be causing this?

Thanks for a great project, we use this in many of our Kubernetes deployments and its been very helpful.

Start at the /metrics handler and the retrieval path in monitoring_collector.go, especially the code around the logged lines 166 and 210. Reproduce the 500 response with appengine.googleapis.com in the prefix list, then identify the failing metric or response conversion and add coverage showing that the handler no longer returns an error.