i come from cricket-monitoring world. an ancient perl frontend to rrd with many features which modern monitoring tools don't reach (yet).
zabbix graphs are the best graphs i've seen since cricket, but imho there's a lack of "visual comfort", particularly regarding timestamps. it's a bit hard to track graphs with timestamp marks such as "02.15 03:34", next "08:50", "14:24", "19:49" and "02.16 01:15". it seems timestamps are calculated from grid marks instead the other way round (first we calc timestamps, then we calc the grid marks to match tstamps).
besides, there's another feature in cricket graphs called "zerotimes": in graphs with period below a week, cricket draws a red vertical line to mark each day change (i.e. 00h00m00s), in graphs below a month, cricket draws a line each monday at 0:00h. in graphs below a year, cricket draws a line each first day of the month, and so.
i think these features would avoid misunderstandings, ease troubleshooting tasks, and in summary, they'd improve zabbix significantly
zabbix graphs are the best graphs i've seen since cricket, but imho there's a lack of "visual comfort", particularly regarding timestamps. it's a bit hard to track graphs with timestamp marks such as "02.15 03:34", next "08:50", "14:24", "19:49" and "02.16 01:15". it seems timestamps are calculated from grid marks instead the other way round (first we calc timestamps, then we calc the grid marks to match tstamps).
besides, there's another feature in cricket graphs called "zerotimes": in graphs with period below a week, cricket draws a red vertical line to mark each day change (i.e. 00h00m00s), in graphs below a month, cricket draws a line each monday at 0:00h. in graphs below a year, cricket draws a line each first day of the month, and so.
i think these features would avoid misunderstandings, ease troubleshooting tasks, and in summary, they'd improve zabbix significantly
)
Comment