Hi , While reviewing Prometheus-related content online, I've come to understand that the rate is calculated as (v2 - v1) / (t2 - t1), and the increase is simply (v2 - v1). Now, considering i have 12 data points (of Gauge type) with a scrape interval of 1 minute, if i apply the rate(metric[2m]) / increase(metric[2m]) or deriv/delta functions, i notice some discrepancies in the results. Example - p1=10 @2024-04-06 13:33:48.666000 p2=10 @2024-04-06 13:34:48.666000 p3=10 @2024-04-06 13:35:48.666000 p4=10 @2024-04-06 13:36:48.666000 p5=20 @2024-04-06 13:37:48.666000 p6=20 @2024-04-06 13:38:48.666000 p7=20 @2024-04-06 13:39:48.666000 p8=20 @2024-04-06 13:40:48.666000 p9=30 @2024-04-06 13:41:48.666000 p10=30 @2024-04-06 13:42:48.666000 p11=30 @2024-04-06 13:43:48.666000 p12=30 @2024-04-06 13:44:48.666000 ....
in that case if i use rate(metric[2m]), and increase(metric[2m]) here: p1: rate = NA, increase=NA ( not enough data points) 13:33:48.666000 p2: rate= 0 (10-10)/120 , increase = 0 (10-10) 13:34:48.666000 - 13:32:48.666000 p3: rate= 0 (10-10)/120 , increase = 0 (10-10) 13:35:48.666000 - 13:33:48.666000 p4: rate= 0 (10-10)/120 , increase = 0 (10-10) 13:36:48.666000 - 13:34:48.666000 p5: rate=0.0833 (20-10)/120 , increase=0 (20-10) 13:37:48.666000 - 13:35:48.666000 p6: rate=0.0833 (20-10)/120 , increase=0 (20-20) 13:38:48.666000 - 13:36:48.666000 p7: rate=0 (20-20)/120 , increase=0 (20-20) 13:39:48.666000 - 13:37:48.666000 p8: rate=0 (20-20)/120 , increase=0 (20-20) 13:40:48.666000 - 13:38:48.666000 p9: rate=0.0833 (30-20)/120 , increase=10 (30-20) 13:41:48.666000 - 13:39:48.666000 p10: rate=0.0833 (30-20)/120 , increase=10 (30-20) 13:42:48.666000 - 13:40:48.666000 p11: rate=0 (30-30)/120 , increase=0 (30-30) 12:43:48.666000 - 12:41:48.666000 p12: rate=0 (30-30)/120 , increase=0 (30-30) 12:44:48.666000 - 12:42:48.666000 But when i tried visualizing the data through the prometheus , [image: increase.png] For instance, for point p5, i expected a rate of 0.0833 and an increase of 10, but Prometheus displayed 0.16 and 20 respectively. Could you please share some insights onto this observation? Regards, P -- 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/d5922e48-29f0-4d81-96ea-6e5765c2c134n%40googlegroups.com.
