This would enable collecting data in addition to what Prometheus collects, and also collect it at a finer granularity (Prometheus metrics are at 30-second granularity; Pbench collects some data at much smaller intervals). Via bench-army-knife it would further be possible to collect data from pods running within VMs (bench-army-knife is currently specific to fio and uperf).
RobertKrawitz
commented
1 year ago
This would enable collecting data in addition to what Prometheus collects, and also collect it at a finer granularity (Prometheus metrics are at 30-second granularity; Pbench collects some data at much smaller intervals). Via bench-army-knife it would further be possible to collect data from pods running within VMs (bench-army-knife is currently specific to fio and uperf).
Reference https://github.com/distributed-system-analysis/pbench