Sounds about right. I do have a number of instances with 6-6.4M time series and they use around 20-35GB of go_memstats_alloc_bytes and around 66-96GB of RSS memory. That's all with GOGC=40. go_memstats_alloc_bytes / prometheus_tsdb_head_series is between 4-8KB on most of them.
On Wednesday, 25 May 2022 at 13:45:09 UTC+1 [email protected] wrote: > Thanks for this info, so if I am understanding correctly. > > prometheus_tsdb_head_series returns 6181063 , when graphing this it almost > looks vertical. > > go_memstats_alloc_bytes / prometheus_tsdb_head_series returns 6741 > > Doing a calculation on that, would come up to roughly 41GB , prior to any > other memory utilization requirements of Prometheus. With that said then > you double it, so looking more like 82GB of memory + allocation for > prometheus overhead. Does this seem about accurate? > > Thanks! > > >>> -- You received this message because you are subscribed to the Google Groups "Prometheus Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/773e7e8f-55a6-40d1-8067-6a2c064b23c1n%40googlegroups.com.
