1latencytop(1M)          System Administration Commands          latencytop(1M)
2
3
4

NAME

6       latencytop  - report latency-related statistics in system and in appli‐
7       cations
8

SYNOPSIS

10       latencytop [-t interval] [-o log_file] [-k log_level]
11            [-f [no]feature,...] [-l log_interval] [-h]
12
13

DESCRIPTION

15       LatencyTOP is an  observability  tool  that  reports  statistics  about
16       latencies  in  the system and in applications. The tool reports statis‐
17       tics about where and what kind of latencies are happening in the system
18       and  in the applications that are running on the system. The statistics
19       then can be used to improve performance throughput of applications  and
20       system, as you remove the identified latencies.
21
22
23       The tool analyzes system activity periodically and displays the data in
24       the output window. Two types of latencies are tracked: an LWP going  in
25       and out of sleep and an LWP spinning order to acquire a synchronization
26       object. The tool uses the Solaris DTrace framework to collect the  sta‐
27       tistics  corresponding to these two scenarios of inactivity of the sys‐
28       tem and application LWPs.
29
30
31       The output window is divided into two sections. An upper part  displays
32       the  system-wide  statistics,  while the lower part displays statistics
33       about individual processes. The user can navigate the list of processes
34       (using  the  left-  and  right-arrow keys) and select the list they are
35       interested in.  The  tool  will  then  display  statistics  about  that
36       selected  process in the lower part of the window. If the t or T key is
37       pressed, the tool displays  the  LWP-specific  view  of  that  selected
38       process.  The t or T key can be used to toggle between the process-view
39       and the thread-view.
40
41
42       During execution, a user can force a refresh of the analysis by  press‐
43       ing  the r or R key. The interval time is restored to the default or to
44       a specified value (if -t was used). To quit the application,  the  user
45       must press the q or Q key.
46

OPTIONS

48       The following options are supported:
49
50       -f, --feature [no]feature1,[no]feature2,...
51
52           Enables/disables  features  in LatencyTOP. Features can be only one
53           of the following:
54
55           [no]filter
56
57               Filter large interruptible latencies, for example,  sleep.  The
58               default is off.
59
60
61           [no]sched
62
63               Monitors sched (PID=0). The default is off.
64
65
66           [no]sobj
67
68               Monitors synchronize objects. The default is on.
69
70
71           [no]low
72
73               Lower  overhead by sampling small latencies. Enabling this fea‐
74               ture will lower CPU utilization by estimating  small  latencies
75               statistically.  Use  it for heavy workloads such as a very busy
76               web server. The default is off.
77
78
79
80       -h
81
82           Displays the command's usage.
83
84
85       -k log_level
86
87           Specifies the level of logging in the log file. Valid values are:
88
89           0    none (default)
90
91
92           1    unknown
93
94
95           2    all
96
97
98
99       -l [log_interval]
100
101           Writes data to the log file every log_interval seconds;  log_inter‐
102           val must be greater than 60.
103
104
105       -o log_file
106
107           Specifies  the  log  file where output will be written. The default
108           log file is /var/log/latencytop.log.
109
110
111       -t interval
112
113           Specifies the interval, in seconds, at which the tool collects sta‐
114           tistics  from the system. The possible values are between 1 and 60;
115           the default is 5 seconds.
116
117

EXAMPLES

119       Example 1 Running the Tool
120
121
122       The following  command  launches  the  tool  with  default  values  for
123       options.
124
125
126         % latencytop
127
128
129
130       Example 2 Setting the Interval
131
132
133       The following command sets the sampling interval to two seconds.
134
135
136         % latencytop -t 2
137
138
139
140       Example 3 Specifying the Log File
141
142
143       The following command sets the log file to /tmp/latencytop.log.
144
145
146         % latencytop -o /tmp/latencytop.log
147
148
149
150       Example 4 Specifying the Log Level
151
152
153       The following command sets the log level to all.
154
155
156         % latencytop -l 2
157
158
159
160       Example 5 Enabling Tracing of Latencies
161
162
163       The  following  command enables the tracing of latencies caused by syn‐
164       chronization objects.
165
166
167         % latencytop -f sobj
168
169
170

EXIT STATUS

172       0
173
174           Successful operation.
175
176
177       1
178
179           An error occurred.
180
181

ATTRIBUTES

183       See attributes(5) for descriptions of the following attributes:
184
185
186
187
188       ┌─────────────────────────────┬─────────────────────────────┐
189       │      ATTRIBUTE TYPE         │      ATTRIBUTE VALUE        │
190       ├─────────────────────────────┼─────────────────────────────┤
191       │Architecture                 │x86, SPARC                   │
192       ├─────────────────────────────┼─────────────────────────────┤
193       │Availability                 │SUNWlatencytop               │
194       ├─────────────────────────────┼─────────────────────────────┤
195       │Interface Stability          │Volatile                     │
196       └─────────────────────────────┴─────────────────────────────┘
197

SEE ALSO

199       dtrace(1M), kstat(1M), attributes(5)
200

USAGE

202       You must have DTrace privileges to run LatencyTOP.
203
204
205
206SunOS 5.11                        29 Sep 2009                   latencytop(1M)
Impressum