Kevin Tung·Oct 29Averaging Ten Host p99s Gave Us 679ms. The Real Fleet p99 Was 783ms.The mean of ten host p99s was 678.5ms. The pooled fleet p99 was 783.3ms. Averaging percentiles is not a fleet rollup.A response icon1A response icon1
Theerut Bunkhanphol·Oct 22Monitoring ที่ดีต้องเริ่มจากอะไรช่วงที่ผ่านมา ผมมีโอกาสทำระบบ Monitoring มากขึ้น และพบว่าคำว่า SLA, SLO และ SLA เป็นคำที่หลายทีมหยิบมาใช้บ่อย…
shahzad bhatti·Jul 20Measuring Availability Properly: Percentiles, Tail Latency, and the Production TrapsObservability is a key part of any infrastructure but I’ve watched teams repeat the same mistakes around measuring availability. For…
InBeyond LocalhostbyThe Speedcraft Lab·Mar 30Our API Had a 6-Second P99 for 8 Months. Nobody Noticed Until a Customer Tweeted.The metric that wasn’t being measured. The alert that was never set up.A response icon2A response icon2
InTechSyncbyDr. Fadi Shaar·Mar 28How Outliers and Percentiles Reveal the Real Performance of Distributed SystemsIf you’re running a microservices-based system, there’s a silent metric that can mislead you more than any other:
Dr. Shouke Wei·Nov 18, 2025Understanding Percentiles in Python: NumPy vs PandasA Practical Guide to np.percentile() and Its Pandas Equivalents
Afifa Hasan Siddiqui·Nov 17, 20250.5 vs 0.9 in PERCENTILE_CONT()PERCENTILE_CONT() is one of those SQL functions that looks simple until you have to explain what values like 0.5 or 0.9 actually mean.A response icon1A response icon1
InJavarevisitedbyArvind Kumar·Sep 11, 2025Understanding Percentiles in Performance Monitoring: A Deep DiveDiscussion about P99, P95, P85 etc..A response icon1A response icon1
Nathan Wollack·Aug 3, 2025Compensation With R — PercentilesWelcome back for the fifth installment of Compensation With R. The series has looked at some common analyses which a compensation…
