Does anyone know if Zabbix is prone to the same problems as cacti with long term graphs? For example Cacti can take 6 x 5 minute average snapshots to determine a 30 minute average which then uses to graph a weekly graph. Then it takes four 30 minute averages to make a 2 hr average to make a monthly graph. Then it does the same calculation to come up with a 1 day snapshot it uses in a yrly graph. This leads to showing averages that are below the actual average. On top of this it calculates the off-business hours bandwidth into the averages. Does anyone know how Zabbix calculates it's long term graphs, does it store each pull it does for a yr and uses that data or does it do some type of summarization?
-
-
ZABBIX stores both detailed history and 1 hour aggregates. You control how long the data should be kept For example, one may keep detailed history one month and 1 hour aggregates for 3 years.
All the values are calculated in a realtime and available immediately for graphing and reporting. No loss of data precision!
Comment