I like that it is easy to setup, has great documentation for all aspects of its functionality (including custom metrics), handles metric data at an aggregate level in addition to granular ones as well as being able to handle large amounts of data.
The UI can be improved with better labels/naming conventions & more intuitive dashboards. Also would love if there was some way to have this run continually without needing manual intervention from us - e.g., via cron job or something similar. We use Prometheus to view our application health across different services by having multiple instances running simultaneously so we don't need to wait until someone logs into each instance manually after work hours or during weekends etc.