US Patent 11321205 Enterprise-scale time series graphite backend infrastructure

An approach is provided in which the approach allocates a set of carbon groups to a set of collection rules. Each of the carbon groups includes a primary carbon and a backup carbon. The approach matches a set of time series metrics data to one of the set of collection rules and, in response to matching the set of time series metrics data to one the collection rules, the approach selects the carbon group allocated to the matched collection rule and directs the set of time series metrics data to the selected carbon group. The set of metrics data is processed by both the primary carbon and the backup carbon in the selected carbon group.