It looks like, while doing  issue 620265 , renaming the alerts for memory.top_10_mobile downstream slipped through.

Probably there are some alerts for memory.top_10_mobile still in place, but using the old metric names.

The new ones we need are:

*/*/memory.top_10_mobile*/memory:*:all_processes:reported_by_os:system_memory:java_heap:proportional_resident_size_avg/*/*
*/*/memory.top_10_mobile*/memory:*:all_processes:reported_by_os:system_memory:native_heap:proportional_resident_size_avg/*/*
*/*/memory.top_10_mobile*/memory:*:all_processes:reported_by_os:system_memory:ashmem:proportional_resident_size_avg/*/*
*/*/memory.top_10_mobile*/memory:*:all_processes:reported_by_os:system_memory:private_dirty_size_avg/*/*
*/*/memory.top_10_mobile*/memory:*:all_processes:reported_by_chrome:gpu:effective_size_avg/*/*
*/*/memory.top_10_mobile*/memory:*:all_processes:reported_by_os:gpu_memory:proportional_resident_size_avg/*/*
*/*/memory.top_10_mobile*/memory:*:all_processes:reported_by_os:system_memory:proportional_resident_size_avg/*/*

Note: I'm assuming that '*' can match an empty string, so this works for both memory.top_10_mobile and memory.top_10_mobile_stress. Hope that is ok.