I tried to follow the template, but really couldn’t fit into it.
We are deploying Grafana on embedded devices, with quite a weak CPU (quad core Cortex-A53 running at 1.5 GHz, think overclocked Pi 3). I did observe decent load when our dashboard refreshes (eyeballing btop output ~2 cpu-seconds).
What I’d like to do is to find a benchmark which is at least somewhat representative of a dashboard’s load, so that I can compare the performance between browsers.
:Hi @jaskij,
Welcome to the 
community support forums !!
We are excited that you joined our OSS community. Please read about some of the FAQs in the community 
I am not sure if the such a tool is available in the OSS version (might be in the Enterprise one)?
Maybe some community-based tools are available…
However, I can only add one point here based on my experience that it all depends on the query which you are pulling from DataSource to get visualised.
If you can filter your required data on the query level then the load will be lesser. Another factor is how many times you are refreshing your data as rapid time intervals will be slower as the dashboard and panel grow.
Lastly, it also depends on the Backend Database as by default Grafana comes with SQLite which does not need any configuration but it is not suitable in the Production environment and better to use either MySQL or PostgreSQL.
Thank you for pointing out that I need to switch database.
As for the benchmark, my searches did not turn up anything, but I was hoping that maybe one of the generic browser benchmarks (such as Speedometer) would be at least somewhat representative.
I see. Thanks for providing more context.
I think the following might be interesting for you to do more in-depth performance benchmarking (also load testing as well).
Grafana Faro for monitoring the Frontend
Grafana Phlare for continuous profiling (have a dedicated forum)
Also coming back to the Backend Database, we have this documentation for High availability so worth to check it out as well.
I hope that all this is helpful to you