When the number of volumes and brick in a cluster is huge then collectd makes a lot of HTTP request to carbon-cache server, For each metrics carbon-cache takes a lot of effort and utilization to push the metrics into graphite. Instead of pushing each metrics as an individual HTTP request, send a list of metrics in a single HTTP call will reduce the CPU and memory usage of carbon-cache.
When the number of volumes and brick in a cluster is huge then collectd makes a lot of HTTP request to carbon-cache server, For each metrics carbon-cache takes a lot of effort and utilization to push the metrics into graphite. Instead of pushing each metrics as an individual HTTP request, send a list of metrics in a single HTTP call will reduce the CPU and memory usage of carbon-cache.