rerun-io/rerun

Documentation on Memory Usage / out-of-memory behavior

jleibs opened this issue · 3 comments

jleibs commented

There's enough nuance here to potentially make for a good how-to guide?

  • Identifying that rerun is using too much memory
  • The UIs involved in inspecting/debugging
  • The flags involved in configuring the behavior
  • Some best practices for reducing overhead
emilk commented

I agree we should improve the documentation here. Currently we have this in rerun_py/USAGE.md:

You can set --memory-limit=16GB to tell the Rerun Viewer to purge older log data when memory use goes above that limit. This is useful for using Rerun in continuous mode, i.e. where you keep logging new data to Rerun forever.

It is still possible to log data faster than the Rerun Viewer can process it, and in those cases you may still run out of memory unless you also set --drop-at-latency=200ms or similar.

emilk commented

We should also make sure that both --memory-limit and --drop-at-latency can be set from the Python SDK and with environment variables:

  • 👍1