From 690328028d8b45a9304a619926c4cc627a009731 Mon Sep 17 00:00:00 2001 From: nemobis Date: Fri, 8 Nov 2024 17:13:05 +0200 Subject: [PATCH] docs: Update RELEX Oy figures (#7491) Update figures for the existing RELEX Oy case study. ### Checklist The following checks are **mandatory**: - [x] My change adheres to [VictoriaMetrics contributing guidelines](https://docs.victoriametrics.com/contributing/). Co-authored-by: Artem Navoiev --- docs/CaseStudies.md | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/docs/CaseStudies.md b/docs/CaseStudies.md index 92fe1c4ba..ea212361f 100644 --- a/docs/CaseStudies.md +++ b/docs/CaseStudies.md @@ -461,20 +461,20 @@ VictoriaMetrics has allowed us to extend data retention for our metrics effortle Across our production VictoriaMetrics clusters, in a 12 months period we go beyond the following figures. -- Active time series: 10M -- Ingestion rate: 300k samples per second -- Total number of datapoints: 4400G -- Data size on disk: 3600 GiB -- Available memory: 320 GiB as seen by kubernetes (160 GiB physical memory for the hosts) -- CPU: 20 cores (AMD EPYC 7763), about 70 % idle +- Active time series: 20M +- Ingestion rate: 1M samples per second (2M before stream aggregation) +- Total number of datapoints: 15T+ +- Data size on disk: 12+ TiB +- Available memory: 1.2 TiB as seen by kubernetes (640 GiB physical memory for the hosts) +- CPU: 80 cores (AMD EPYC 7763), about 50 % idle - Retention period: ~1 year -- Churn rate: 6M new time series per day (monthly average) +- Churn rate: 16M new time series per day (monthly average) - Query rate: - - `/api/v1/query_range`: 10 queries per second + - `/api/v1/query_range`: 15 queries per second - `/api/v1/query`: 10 queries per second - Query duration for `/api/v1/query_range` (weekly mean): - - 99th percentile: 700 ms - - median: 10 ms + - 99th percentile: 1.5 s + - median: 15 ms ## Roblox