[10:22:30] <gilles>	 reworking the survey dashboard revealed something really cool: https://twitter.com/MonsieurPerf/status/1314148751183929344
[12:04:31] <phedenskog>	 gilles: blog post!
[12:05:52] <gilles>	 Yeah sure, when I have some time. I’ll need to pull the historical per-country nav timing data to plot load time by wiki
[12:10:30] <phedenskog>	 Could be a cross-post in the perf calendar
[13:23:05] <gilles>	 good idea, I can save it for that
[13:23:15] <gilles>	 I probably won't have anything else this good by then
[19:13:49] <gilles>	 https://blog.bearer.sh/how-rust-lets-us-monitor-30k-api-calls-min/
[21:05:52] <Krinkle>	 hm.. having trouble following the logic there. I'm sure it made sense for them, but the details are a bit thin. They switched programming languages for a tiny in-take application processing a few K req/min, because of a memory retention issue in an upstream library?
[21:56:40] <gilles>	 Yeah it sounds like the original had severe problems, it’s not a lot of traffic nor data at all