Monitoring Latency and Transfer Rate per Aggregate

Our new check PerfAggregate  measures various performance counters per aggregate. Currently, the following counters can be monitored:

  • aggr_throughput (B/s)
  • latency (µs)
  • read_data (B/s)
  • total_transfers (/s)
  • total_transfers_hdd (/s)
  • total_transfers_ssd (/s)
  • transmit_failure (-)
  • user_read_blocks (/s)
  • user_read_latency (us)
  • user_write_blocks (/s)
  • user_write_latency (us)
  • write_data (B/s)

We are open to your suggestions on additional counters that we could add. A description for each counter of a filer can be displayed using the getter and respective switch ‑‑discover=counters.``` $ ./ ‑H your_filer ‑o aggregate ‑‑discover=counters

Filer Level

In order to monitor these values on a filer level, we've also included the switch `‑‑math=total|average`, which is already in use in PerfVolume. This way, one can calculate sums and averages for all aggregates. Two examples can be found on our [help page for PerfAggregate]( This plugin is now available and is part of our **unstable (beta) release 3.9.2\_08.** We are planing to release the stable release 3.10.0 in mid-December.

Checking the Metro-Cluster Configuration
Release Candidate 1 for Version 3.10.0 available