Language detection causing excessive CPU usage
It was brought to our attention by a customer (https://gitlab.zendesk.com/agent/tickets/115860 internal-only link) that the git-linguist process is consuming more CPU than expected. Specifically, it is processing a large ".ktr" xml file, taking longer than 40 minutes, and consuming 100% of the CPU.
I'd like to recommend that we could implement some sort of nice usage or a timeout so this is less likely.
It seems performance of language detection has been discussed before: