Hey,
I've been experimenting a lot with the trendstl function in Zabbix 6.0 and have made a custom item which runs my python script that basically either sends the value 50 or value 100, based on the hour. It is set to switch each hour with the following statement: hour % 2 == 0 : print(50) else print(100).
I've done this to create seasonal data. See the following picture for the trendline.
​When using the following trigger: trendstl(<dummy_data>,120h:now/h,96h,24h)>=(1/96), I get this result: 1 >= (1/96), which basically means that it detected a 100% anomaly rate.
This is the total opposite than my expectation because I expect to see 0 anomalies in this trend.
Any thoughts on why this might be happening?
Greetings,
machinelearningdingetjes
I've been experimenting a lot with the trendstl function in Zabbix 6.0 and have made a custom item which runs my python script that basically either sends the value 50 or value 100, based on the hour. It is set to switch each hour with the following statement: hour % 2 == 0 : print(50) else print(100).
I've done this to create seasonal data. See the following picture for the trendline.
​When using the following trigger: trendstl(<dummy_data>,120h:now/h,96h,24h)>=(1/96), I get this result: 1 >= (1/96), which basically means that it detected a 100% anomaly rate.
This is the total opposite than my expectation because I expect to see 0 anomalies in this trend.
Any thoughts on why this might be happening?
Greetings,
machinelearningdingetjes