Hi,
I am wondering if the following request will be possible to implement or not. The idea is that we like to see a more robust/resilient xymon client.
Scenario: if the xymon server ( or group of servers ) is down due to network outage or system failure, all the client's data/status/metrics will be missing during the down time window.
Suggestion: If there is a backlog mechanism, the client can still deliver the data to hobbit server. Although the status alerts will be meaningless and should be ignored, the system statistcis ( netstat/ifstat/load average/vmstat/filesystem usage... etc ) will still be logged to the rrd. The key to such implementation is to have hobbitd_rrd log the data with the appropriate timestamp from the client incoming data, instead of the current system time, if the data is sent with a backlog flag. Otherwise if there is no backlog flag, use the system time stamp to update the rrd.
I'd appreciate if some can tell me if such feature can be achieved by using current xymon releases ( 4.2 to 4.3 )
Edmond