Closed matthewberry closed 6 months ago
Both staging and production are now using the GPU, without limits :+1:
This means that if someone were to hit both instances with a very large PDF at the same time, we might see some slow down. I think the odds of this happening are low, since we should be the only folks using staging (and even then, only for light testing)
I think setting
resource.limits
on thechemscraper-services
deployment prevents other containers from mounting and using the GPU.Removed this section for testing, and so far things seem to be working: https://github.com/moleculemaker/chemscraper-helm-chart/pull/3/files#diff-7792a5fea7287a039634c3c2ee77f1b23d7fcd8b8f6e225446a9e5456dd77a38R40-R43