When developing crash-reporter, or even simply running a dev or test image, there's a few settings to be aware of that control/override runtime behavior.
If testing via ssh, or you otherwise skipped setting up consent, you can opt-in to crash collection by running:
# metrics_client -C
By default, coredumps will be removed after creating minidumps. You can touch /root/.leave_core to change that behavior.
Similarly, if you want Chrome coredumps to be retained, you can touch /mnt/stateful_partition/etc/collect_chrome_crashes.
Crash uploading is disabled on test images. You can force them to be uploaded by running:
# crash_sender --dev --max_spread_time=0
The --max_spread_time=0 option is to make crash_sender upload right away. Otherwise it'll sleep a random amount of time (up to 10 minutes) between reports.
Crash reports uploaded with the --dev flag will be sent to the staging crash server which can be access at http://go/crash-staging.