I'm using collectd with a custom writer plugin. The writer reports to servers in a different data center and occasionally we have network connection issues between the data centers. Collectd continues to work but falls behind. After the network connections return to normal collectd continues to report at the same rate as specified by Interval in the config and never catches up.
Is there something I need to do in the writer plugin so that collectd will report more often or send more data to flush the queue and get caught back up?
The problem was that more metrics were being reported than could be send in the current 1 minute interval. The network issues just made the problem worse. The solution is to change the interval to 5 minutes.