def test_format_num(self):
"""test various numbers with separator"""
self.assertEqual(humanize.format_num(999), "999")
self.assertEqual(humanize.format_num(1000), "1,000")
self.assertEqual(humanize.format_num(1000000), "1,000,000")
self.assertEqual(humanize.format_num(999.4), "999")
self.assertEqual(humanize.format_num(1000.4), "1,000")
self.assertEqual(humanize.format_num(1000000.4), "1,000,000")
def test_format_num():
"""test various numbers with separator"""
assert humanize.format_num(999) == "999"
assert humanize.format_num(1000) == "1,000"
assert humanize.format_num(1000000) == "1,000,000"
assert humanize.format_num(999.4) == "999"
assert humanize.format_num(1000.4) == "1,000"
assert humanize.format_num(1000000.4) == "1,000,000"
def format_table_stat(stat, val_suffix=""):
"""display most value read in a human friendly
format with the responsible table and node"""
if not stat or len(stat) < 3:
return NA
val_suffix_str = ""
if val_suffix:
val_suffix_str = "%s " % val_suffix
node = stat[0]
table = stat[1]
val = stat[2]
if node and table and val:
return "%s %s(%s %s)" % (humanize.format_num(val), val_suffix_str,
table, node)
return NA
def print_summary(self):
""" prints a summary report """
self.analyze()
summary = Summary(self.nodes)
print("%s version: %s" % (self.command_name, VERSION))
print('')
print("Summary (%s lines)" % format_num(summary.lines))
print("Summary (%s skipped lines)" % format_num(summary.skipped_lines))
print('')
print("dse versions: %s" % (set(summary.versions) or 'unknown'))
print("cassandra versions: %s" %
(set(summary.cassandra_versions) or 'unknown'))
print("first log time: %s" % summary.start)
print("last log time: %s" % summary.end)
if not self.dumps_analyzed:
print("Nothing found!")
return
print("duration: %s" %
format_seconds(int(summary.duration.total_seconds())))
print("total stages analyzed: %s" % self.dumps_analyzed)
print("total nodes analyzed: %s" % len(summary.nodes))
pauses = summary.get_pauses()
if pauses:
print('')
percentiles = []
percentiles.append(get_percentile_headers("GC pauses"))
percentiles.append(["", "---", "---", "---", "---", "---", "---"])
percentiles.append(get_percentiles("ms", pauses, strformat="%i"))
pad_table(percentiles, min_width=11, extra_pad=2)
for line in percentiles:
print("".join(line))
print("total GC events: %s" % len(pauses))
print('')
ops = summary.get_busiest_tables('ops')[:5]
if ops:
print("busiest tables by ops across all nodes")
print('-' * 30)
for onode, (oname, num) in ops:
print("* %s: %s: %s" % (onode, oname, num))
print('')
print("busiest table by data across all nodes")
print('-' * 30)
for dnode, (dname, data) in summary.get_busiest_tables('data')[:5]:
print("* %s: %s: %s" % (dnode, dname, data))
print('')
print("busiest stages across all nodes")
print('-' * 30)
data = []
for (name, status, stage, value) in summary.get_busiest_stages():
data.append(
["* %s %s:" % (stage, status),
str(value),
"(%s)" % name])
pad_table(data, extra_pad=2)
for line in data:
print("".join(line))
pending = summary.get_stages_in('pending')
data = []
if pending:
data.append([])
self.__print_stages('pending', pending, data)
delayed = summary.get_stages_in('local backpressure')
if delayed:
data.append([])
self.__print_backpressure(delayed, data)
data.append([])
pad_table(data, extra_pad=2)
for line in data:
print("".join(line))
self.__print_recs()
def __repr__(self):
return "%s ops / %s data" % (format_num(
self.ops), format_bytes(self.data))
def print_histogram(self):
""" print histogram report, analyzing if necessary """
self.analyze()
print("%s version: %s" % (self.command_name, VERSION))
print('')
print("Histogram")
print('')
if not self.nodes:
print("Nothing found!")
return
for name, node in self.nodes.items():
print(name)
print('-' * 60)
print("%s lines" % format_num(node.lines))
print("%s skipped lines" % format_num(node.skipped_lines))
print("dse version: %s" % (node.version or 'unknown'))
print("cassandra version: %s" %
(node.cassandra_version or 'unknown'))
print("log start time: %s" % node.start)
print("log end time: %s" % node.end)
if not node.dumps_analyzed:
print("Nothing found!")
continue
print("duration: %s" %
format_seconds(int(node.duration().total_seconds())))
print("stages analyzed: %s" % node.dumps_analyzed)
if node.pauses:
percentiles = [[]]
percentiles.append(get_percentile_headers("GC pauses"))
header = [""]
header.extend(["---" for i in range(6)])
percentiles.append(header)
percentiles.append(
get_percentiles("ms", node.pauses, strformat="%i"))
pad_table(percentiles, min_width=11, extra_pad=2)
for line in percentiles:
print("".join(line))
print("total GC events: %s" % len(node.pauses))
print('')
ops = node.get_busiest_tables('ops')[:5]
if ops:
print("busiest tables (ops)")
print('-' * 30)
nlen = max(len(o[0]) for o in ops)
for n, t in ops:
print(n.ljust(nlen), t)
data = node.get_busiest_tables('data')[:5]
print("busiest tables (data)")
print('-' * 30)
nlen = max(len(d[0]) for d in data)
for n, t in data:
print(n.ljust(nlen), t)
print('')
if node.stages:
percentiles = []
percentiles.append(get_percentile_headers('stages'))
percentiles.append([])
for status, stage in node.stages.items():
header = [status.upper()]
header.extend("" for i in range(6))
percentiles.append(header)
for tpname, vals in sorted(stage.items(),
key=lambda tpool: max(tpool[1]),
reverse=True):
percentiles.append(
get_percentiles(tpname, vals, strformat="%i"))
percentiles.append([])
pad_table(percentiles, extra_pad=2)
for line in percentiles:
print("".join(line))
print('')
self.__print_recs()