The process that kept dying: A memory leak murder mystery
lukedeniston.comThe site is broken right now, but I read this when it was first submitted some time ago.
Contrary to what the title would suggest, it's about finding a mundane JS memory leak in moment.js by attaching the chrome inspector to node. There's no out-of-the-ordinary tale here and there's certainly little mystery.
The article might be useful if you've never done it before and need some pointers.
I'd like to read this article but keep getting ERR_TOO_MANY_REDIRECTS. Sorry, your site seems broken.
Thank you so much for letting me know. Tried out cloudflare this morning, rolling it back.
Is anyone else getting a redirect loop when trying the link? At first I was thinking it had to do with my Firefox settings to force HTTPS but even just:
curl -L lukedeniston.com/memory-leak-mystery
> curl: (47) Maximum (50) redirects followed
isn't working.
It's me... tried out cloudflare this morning. Bad timing.
>The NODE_OPTIONS were only applying to the yarn process, not the underlying Next.js server that was being invoked by yarn.
Doesn't make sense. All child processes should get that env var.
In my team we actually had to do performance test every time we update our libraries, and never auto update.
We certainly do performance tests, and we also don't auto-update. The issue here was it was a slow leak, so it took time and significant load for the pattern of errors to present itself.