Whilst Prometheus is a popular solution for fordwarding metrics this piece of open-source software can often prove tricky when it requires scaling to be suitable for enterprise use cases.
One common issue with using Prometheus is that gaining a global view of performance metrics can be difficult due having to switch between seperate Prometheus servers running in different regions.
This leads to the user also having to deploy multiple instances of Grafana or having to list different Prometheus servers as separate data sources within Grafana.
Whilst switching between different regions isn't too tedious, users that want a global view will find many shortcomings with having to use separate Prometheus servers for each region when they want to quickly see Global CPU usage or conduct capacity planning.
This is where additional tools such as Cortex or a manual process such as Federation bring their own challenges to maintain as part of your extended monitoring infrastructure.
If you are looking for a solution that provides horizontal scale and highly available Prometheus, then Logit.io's Prometheus as a Service gives you a global view of all of your metrics with zero gaps in data (due to bypassing the need to federate) as well as a fully managed service for hosted Grafana.