measure-rgw-latency.py

# assuming we are running a vstart cluster
import subprocess
import time
import sys

def system(cmd):
  output = subprocess.check_output(cmd, shell=True).decode(sys.stdout.encoding).strip()
  return output

prev_rgw_sum = 0
prev_count = 0

while True:
  pid = system('pgrep radosgw | head -1')
  cpu = system('top -p '+pid+' -b -n 1 | ag radosgw | awk \'{print $9}\'')
  qlen = system('./bin/ceph --admin-daemon out/radosgw.8000.asok perf dump 2>/dev/null | jq .rgw.qlen')
  obj_lat_sum = float(system('./bin/ceph --admin-daemon out/radosgw.8000.asok perf dump 2>/dev/null | jq .objecter.op_latency.sum'))
  obj_inflight = int(system('./bin/ceph --admin-daemon out/radosgw.8000.asok perf dump 2>/dev/null | jq .objecter.op_inflight'))
  lat_count = int(system('./bin/ceph --admin-daemon out/radosgw.8000.asok perf dump 2>/dev/null | jq .rgw.put_initial_lat.avgcount'))
  lat_sum = float(system('./bin/ceph --admin-daemon out/radosgw.8000.asok perf dump 2>/dev/null | jq .rgw.put_initial_lat.sum'))

  rgw_sum = lat_sum - obj_lat_sum
  diff_sum = rgw_sum - prev_rgw_sum
  diff_count = lat_count - prev_count
  if diff_count == 0:
    latency = 0.0
  else:
    latency = diff_sum/diff_count

  prev_rgw_sum = rgw_sum
  prev_count = lat_count

  print(qlen, '%.4f'%latency, cpu,  obj_inflight)
  time.sleep(5)