20112016 Ericsson AB. All Rights Reserved. Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. Observer observer_ug.xml
Introduction

Observer is a graphical tool for observing the characteristics of Erlang systems. Observer displays system information, application supervisor trees, process information, ETS tables, Mnesia tables and contains a front end for Erlang tracing.

Getting Started

Run Observer from a standalone node to minimize the impact of the system being observed.

Example:

% erl -sname observer -hidden -setcookie MyCookie -run observer

Select the node to observe with menu Nodes. Menu View > Refresh interval controls how often the view is to be updated. The refresh interval is set per viewer so you can have different settings for each viewer. To minimize the system impact, only the active viewer is updated. Other views are updated when activated.

The mouse buttons behave as expected. Use left-click to select objects, right-click to get a menu with the most used options, and double-click to display information about the selected object. In most viewers with many columns, you can change the sort order by left-clicking the column header.

System Tab

Tab System displays general information about the active Erlang node and its runtime system, such as build configuration, system capabilities, and overall use statistics.

Load Charts Tab

Tab Load Charts displays graphs of the current resource use on the active Erlang node.

Graph Scheduler Utilization shows scheduler use per scheduler, where each scheduler use has a unique color.

Graph Memory Usage shows the total memory use and per memory category use, where each category has a unique color. The categories are as follows:

Total

The sum of all memory categories.

Processes

The sum of all process memory used.

Atom

The size used by the atom table.

Binary

The sum of all off-heap binaries allocated.

Code

The memory allocated for code storage.

Ets

The used memory for all ETS tables.

Graph IO Usage shows the current I/O load on the system.

Memory Allocators Tab

Tab Memory Allocators displays detailed information of the carrier size and current memory carriers. For details about memory carriers, see module erts_alloc in application ERTS.

Applications Tab

Tab Applications presents application information. Select an application in the left list to display its supervisor tree. The right-click options in the tree are as follows:

Process info

Opens a detailed information window on the selected process, including the following:

Process Information

Shows the process information.

Messages

Shows the process messages.

Dictionary

Shows the process dictionary.

Stack Trace

Shows the process current stack trace.

State

Shows the process state.

Log

If enabled and available, shows the process SASL log entries.

Trace process

Adds the selected process identifier to tab Trace Overview plus the node that the process resides on.

Trace named process

Adds the registered name of the process. This can be useful when tracing on many nodes, as processes with that name are then traced on all traced nodes.

Trace process tree

Adds the selected process and all processes below, right of it, to tab Trace Overview.

Trace named process tree

Adds the selected process and all processes below, right of it, to tab Trace Overview.

Processes Tab

Tab Processes lists process information in columns. For each process the following information is displayed:

Pid

The process identifier.

Reds

The number of reductions executed on the process. This can be presented as accumulated values or as values since the last update.

Memory

The size of the process, in bytes, obtained by a call to process_info(Pid,memory).

MsgQ

The length of the message queue for the process.

Option Process info opens a detailed information window on the process under the mouse pointer, including the following:

Process Information

Shows the process information.

Messages

Shows the process messages.

Dictionary

Shows the process dictionary.

Stack Trace

Shows the process current stack trace.

State

Shows the process state.

Log

If enabled and available, shows the process SASL log entries.

Log requires application SASL to be started on the observed node, with log_mf_h as log handler. The Observed node must be Erlang/OTP R16B02 or higher. The rb server must not be started on the observed node when clicking menu Log > Toggle log view. The rb server is stopped on the observed node when exiting or changing the observed node.

Option Trace selected processes adds the selected process identifiers to tab Trace Overview plus the node that the processes reside on.

Option Trace selected processes by name adds the registered name of the processes. This can be useful when tracing is done on many nodes, as processes with that name are then traced on all traced nodes.

Option Kill process brutally kills the processes under the mouse pointer by sending an exit signal with reason kill.

Ports Tab

Tab Ports lists port information in columns. For each port the following information is displayed:

Id

The port identifier.

Connected

The process identifier for the process that owns the port.

Name

The registered name of the port, if any.

Controls

The name of the command set by erlang:open_port/2.

Slot

The internal index of the port.

Option Port info opens a detailed information window for the port under the mouse pointer. In addition to the information above, it also shows links and monitors.

Option Trace selected ports adds the selected port identifiers, and the nodes that the ports reside on, to tab Trace Overview.

Option Trace selected ports by name adds the registered name of the port to tab Trace Overview. This can be useful when tracing is done on many nodes, as ports with that name are then traced on all traced nodes.

Option Close executes erlang:port_close/1 on the port under the mouse pointer.

Table Viewer Tab

Tab Table Viewer lists tables. By default, ETS tables are displayed whereas unreadable private ETS tables and tables created by OTP applications are not diplayed. Use menu View to view "system" ETS tables, unreadable ETS tables, or Mnesia tables.

Double-click to view the table content, or right-click and select option Show Table Content. To view table information, select the table and activate menu View > Table information, or right-click and select option Table info.

You can use regular expressions and search for objects, and edit or delete them.

Trace Overview Tab

Tab Trace Overview handles tracing. Trace by selecting the processes or ports to be traced and how to trace them. For processes, you can trace messages, function calls, scheduling, garbage collections, and process-related events such as spawn, exit, and many others. For ports, you can trace messages, scheduling and port-related events.

To trace function calls, you also need to set up trace patterns. Trace patterns select the function calls to be traced. The number of traced function calls can be further reduced with match specifications. Match specifications can also be used to trigger more information in the trace messages.

You can also set match specifications on messages. By default, if tracing messages, all messages sent and/or received by the process or port are traced. Match specifications can be used to reduce the number of traced messages and/or to trigger more information in the trace messages.

Trace patterns only apply to the traced processes and ports.

Processes are added from the Applications or Processes tabs. Ports are added from the Ports tab. A special new identifier, meaning all processes, or ports, started after trace start, can be added with buttons Add 'new' Processes and Add 'new' Ports, respecively.

When adding processes or ports, a window with trace options is displayed. The chosen options are set for the selected processes/ports. To change the options, right-click the process or port and select Edit process options. To remove a process or port from the list, right-click and select Remove process or Remove port, respectively.

Processes and ports added by process/port identifiers add the nodes these processes/ports reside on in the node list. More nodes can be added by clicking button Add Nodes, or by right-clicking in the Nodes list and select Add Nodes. To remove nodes, select them, then right-click and choose Remove nodes.

If function calls are traced, trace patterns must be added by clicking button Add Trace Pattern. Select a module, function(s), and a match specification. If no functions are selected, all functions in the module are traced.

Trace patterns can also be added for traced messages. Click button Add Trace Pattern and select Messages sent or Messages received, and a match specification.

A few basic match specifications are provided in the tool, and you can provide your own match specifications. The syntax of match specifications is described in the ERTS User's Guide. To simplify the writing of a match specification, they can also be written as fun/1. For details, see module ms_transform in application STDLIB.

Click button Start Trace to start the trace. By default, trace output is written to a new window. Tracing is stopped when the window is closed, or when clicking button Stop Trace. Trace output can be changed with menu Options > Output. The trace settings, including match specifications, can be saved to, or loaded from, a file.

For details about tracing, see module dbg in application Runtime_Tools and in section "Match specifications in Erlang" in ERTS User's Guide and in module ms_transform in application STDLIB.