dotnet-trace performance analysis utility

This article applies to: ✔️ dotnet-trace 3.0.47001 and later versions

Install

There are two ways to download and install dotnet-trace:

Note

To use dotnet-trace on an x86 app, you need a corresponding x86 version of the tool.

Synopsis

dotnet-trace [-h, --help] [--version] <command>

Description

The dotnet-trace tool:

  • Is a cross-platform .NET Core tool.
  • Enables the collection of .NET Core traces of a running process without a native profiler.
  • Is built on EventPipe of the .NET Core runtime.
  • Delivers the same experience on Windows, Linux, or macOS.

Options

  • -h|--help

    Shows command-line help.

  • --version

    Displays the version of the dotnet-trace utility.

Commands

Command
dotnet-trace collect
dotnet-trace convert
dotnet-trace ps
dotnet-trace list-profiles
dotnet-trace report

dotnet-trace collect

Collects a diagnostic trace from a running process or launches a child process and traces it (.NET 5+ only). To have the tool run a child process and trace it from its startup, append -- to the collect command.

Synopsis

dotnet-trace collect [--buffersize <size>] [--clreventlevel <clreventlevel>] [--clrevents <clrevents>]
    [--format <Chromium|NetTrace|Speedscope>] [-h|--help]
    [-n, --name <name>] [--diagnostic-port] [-o|--output <trace-file-path>] [-p|--process-id <pid>]
    [--profile <profile-name>] [--providers <list-of-comma-separated-providers>]
    [--show-child-io]
    [-- <command>] (for target applications running .NET 5 or later)

Options

  • --buffersize <size>

    Sets the size of the in-memory circular buffer, in megabytes. Default 256 MB.

    Note

    If the target process writes events too frequently, it can overflow this buffer and some events might be dropped. If too many events are getting dropped, increase the buffer size to see if the number of dropped events reduces. If the number of dropped events does not decrease with a larger buffer size, it may be due to a slow reader preventing the target process' buffers from being flushed.

  • --clreventlevel <clreventlevel>

    Verbosity of CLR events to be emitted.

  • --clrevents <clrevents>

    A list of CLR runtime provider keywords to enable separated by + signs. This is a simple mapping that lets you specify event keywords via string aliases rather than their hex values. For example, dotnet-trace collect --providers Microsoft-Windows-DotNETRuntime:3:4 requests the same set of events as dotnet-trace collect --clrevents gc+gchandle --clreventlevel informational. The table below shows the list of available keywords:

    Keyword String Alias Keyword Hex Value
    gc 0x1
    gchandle 0x2
    fusion 0x4
    loader 0x8
    jit 0x10
    ngen 0x20
    startenumeration 0x40
    endenumeration 0x80
    security 0x400
    appdomainresourcemanagement 0x800
    jittracing 0x1000
    interop 0x2000
    contention 0x4000
    exception 0x8000
    threading 0x10000
    jittedmethodiltonativemap 0x20000
    overrideandsuppressngenevents 0x40000
    type 0x80000
    gcheapdump 0x100000
    gcsampledobjectallocationhigh 0x200000
    gcheapsurvivalandmovement 0x400000
    gcheapcollect 0x800000
    gcheapandtypenames 0x1000000
    gcsampledobjectallocationlow 0x2000000
    perftrack 0x20000000
    stack 0x40000000
    threadtransfer 0x80000000
    debugger 0x100000000
    monitoring 0x200000000
    codesymbols 0x400000000
    eventsource 0x800000000
    compilation 0x1000000000
    compilationdiagnostic 0x2000000000
    methoddiagnostic 0x4000000000
    typediagnostic 0x8000000000

    You can read about the CLR provider more in detail on the .NET runtime provider reference documentation.

  • --format {Chromium|NetTrace|Speedscope}

    Sets the output format for the trace file conversion. The default is NetTrace.

  • -n, --name <name>

    The name of the process to collect the trace from.

  • --diagnostic-port <path-to-port>

    The name of the diagnostic port to create. See Use diagnostic port to collect a trace from app startup to learn how to use this option to collect a trace from app startup.

  • -o|--output <trace-file-path>

    The output path for the collected trace data. If not specified, it defaults to trace.nettrace.

  • -p|--process-id <PID>

    The process ID to collect the trace from.

  • --profile <profile-name>

    A named pre-defined set of provider configurations that allows common tracing scenarios to be specified succinctly. The following profiles are available:

Profile Description
cpu-sampling Useful for tracking CPU usage and general .NET runtime information. This is the default option if no profile or providers are specified.
gc-verbose Tracks GC collections and samples object allocations.
gc-collect Tracks GC collections only at very low overhead.
  • --providers <list-of-comma-separated-providers>

    A comma-separated list of EventPipe providers to be enabled. These providers supplement any providers implied by --profile <profile-name>. If there's any inconsistency for a particular provider, this configuration takes precedence over the implicit configuration from the profile.

    This list of providers is in the form:

    • Provider[,Provider]
    • Provider is in the form: KnownProviderName[:Flags[:Level][:KeyValueArgs]].
    • KeyValueArgs is in the form: [key1=value1][;key2=value2].

    To learn more about some of the well-known providers in .NET, refer to Well-known Event Providers.

  • -- <command> (for target applications running .NET 5 only)

    After the collection configuration parameters, the user can append -- followed by a command to start a .NET application with at least a 5.0 runtime. This may be helpful when diagnosing issues that happen early in the process, such as startup performance issue or assembly loader and binder errors.

    Note

    Using this option monitors the first .NET 5 process that communicates back to the tool, which means if your command launches multiple .NET applications, it will only collect the first app. Therefore, it is recommended you use this option on self-contained applications, or using the dotnet exec <app.dll> option.

  • --show-child-io

    Shows the input and output streams of a launched child process in the current console.

Note

  • Stopping the trace may take a long time (up to minutes) for large applications. The runtime needs to send over the type cache for all managed code that was captured in the trace.
  • On Linux and macOS, this command expects the target application and dotnet-trace to share the same TMPDIR environment variable. Otherwise, the command will time out.
  • To collect a trace using dotnet-trace, it needs to be run as the same user as the user running the target process or as root. Otherwise, the tool will fail to establish a connection with the target process.
  • If you see an error message similar to: [ERROR] System.ComponentModel.Win32Exception (299): A 32 bit processes cannot access modules of a 64 bit process., you are trying to use a version of dotnet-trace that has mismatched bitness against the target process. Make sure to download the correct bitness of the tool in the install link.
  • If you experience an unhandled exception while running dotnet-trace collect, this results in a broken trace. If finding the root cause of the exception is your priority, navigate to Collect dumps on crash. As a result of the crash in the program, the trace is truncated when the runtime rips apart to prevent breaking other parts of the program. Even though the trace is broken, you can still open it to see what happened leading up to the failure. However, it will be missing Rundown information (this happens at the end of a trace) so stacks might be unresolved (depending on what providers were turned on). Open the trace by executing PerfView with the /ContinueOnError flag at the command line. The logs will also contain the location the exception was fired.

dotnet-trace convert

Converts nettrace traces to alternate formats for use with alternate trace analysis tools.

Synopsis

dotnet-trace convert [<input-filename>] [--format <Chromium|NetTrace|Speedscope>] [-h|--help] [-o|--output <output-filename>]

Arguments

  • <input-filename>

    Input trace file to be converted. Defaults to trace.nettrace.

Options

  • --format <Chromium|NetTrace|Speedscope>

    Sets the output format for the trace file conversion.

  • -o|--output <output-filename>

    Output filename. Extension of target format will be added.

Note

Converting nettrace files to chromium or speedscope files is irreversible. speedscope and chromium files don't have all the information necessary to reconstruct nettrace files. However, the convert command preserves the original nettrace file, so don't delete this file if you plan to open it in the future.

dotnet-trace ps

Lists the dotnet processes that traces can be collected from. dotnet-trace 6.0.320703 and later, also display the command-line arguments that each process was started with, if available.

Synopsis

dotnet-trace ps [-h|--help]

Example

Suppose you start a long-running app using the command dotnet run --configuration Release. In another window, you run the dotnet-trace ps command. The output you'll see is as follows. The command-line arguments, if available, are shown in dotnet-trace version 6.0.320703 and later.

> dotnet-trace ps
  
  21932 dotnet     C:\Program Files\dotnet\dotnet.exe   run --configuration Release
  36656 dotnet     C:\Program Files\dotnet\dotnet.exe

dotnet-trace list-profiles

Lists pre-built tracing profiles with a description of what providers and filters are in each profile.

Synopsis

dotnet-trace list-profiles [-h|--help]

dotnet-trace report

Creates a report into stdout from a previously generated trace.

Synopsis

dotnet-trace report [-h|--help] <tracefile> [command]

Arguments

  • <tracefile>

    The file path for the trace being analyzed.

Commands

dotnet-trace report topN

Finds the top N methods that have been on the callstack the longest.

Synopsis
dotnet-trace report <tracefile> topN [-n|--number <n>] [--inclusive] [-v|--verbose] [-h|--help]
Options
  • -n|--number <n>

Gives the top N methods on the callstack.

  • --inclusive

Output the top N methods based on inclusive time. If not specified, exclusive time is used by default.

  • -v|--verbose

Output the parameters of each method in full. If not specified, parameters will be truncated.

Collect a trace with dotnet-trace

To collect traces using dotnet-trace:

  • Get the process identifier (PID) of the .NET Core application to collect traces from.

    • On Windows, you can use Task Manager or the tasklist command, for example.
    • On Linux, for example, the ps command.
    • dotnet-trace ps
  • Run the following command:

    dotnet-trace collect --process-id <PID>
    

    The preceding command generates output similar to the following:

    Press <Enter> to exit...
    Connecting to process: <Full-Path-To-Process-Being-Profiled>/dotnet.exe
    Collecting to file: <Full-Path-To-Trace>/trace.nettrace
    Session Id: <SessionId>
    Recording trace 721.025 (KB)
    
  • Stop collection by pressing the <Enter> key. dotnet-trace will finish logging events to the trace.nettrace file.

Launch a child application and collect a trace from its startup using dotnet-trace

Important

This works for apps running .NET 5 or later only.

Sometimes it may be useful to collect a trace of a process from its startup. For apps running .NET 5 or later, it is possible to do this by using dotnet-trace.

This will launch hello.exe with arg1 and arg2 as its command-line arguments and collect a trace from its runtime startup:

dotnet-trace collect -- hello.exe arg1 arg2

The preceding command generates output similar to the following:

No profile or providers specified, defaulting to trace profile 'cpu-sampling'

Provider Name                           Keywords            Level               Enabled By
Microsoft-DotNETCore-SampleProfiler     0x0000F00000000000  Informational(4)    --profile
Microsoft-Windows-DotNETRuntime         0x00000014C14FCCBD  Informational(4)    --profile

Process        : E:\temp\gcperfsim\bin\Debug\net5.0\gcperfsim.exe
Output File    : E:\temp\gcperfsim\trace.nettrace


[00:00:00:05]   Recording trace 122.244  (KB)
Press <Enter> or <Ctrl+C> to exit...

You can stop collecting the trace by pressing <Enter> or <Ctrl + C> key. Doing this will also exit hello.exe.

Note

Launching hello.exe via dotnet-trace will redirect its input/output and you won't be able to interact with it on the console by default. Use the --show-child-io switch to interact with its stdin/stdout. Exiting the tool via CTRL+C or SIGTERM will safely end both the tool and the child process. If the child process exits before the tool, the tool will exit as well and the trace should be safely viewable.

Use diagnostic port to collect a trace from app startup

Important

This works for apps running .NET 5 or later only.

Diagnostic port is a runtime feature added in .NET 5 that allows you to start tracing from app startup. To do this using dotnet-trace, you can either use dotnet-trace collect -- <command> as described in the examples above, or use the --diagnostic-port option.

Using dotnet-trace <collect|monitor> -- <command> to launch the application as a child process is the simplest way to quickly trace the application from its startup.

However, when you want to gain a finer control over the lifetime of the app being traced (for example, monitor the app for the first 10 minutes only and continue executing) or if you need to interact with the app using the CLI, using --diagnostic-port option allows you to control both the target app being monitored and dotnet-trace.

  1. The command below makes dotnet-trace create a diagnostics socket named myport.sock and wait for a connection.

    dotnet-trace collect --diagnostic-port myport.sock
    

    Output:

    Waiting for connection on myport.sock
    Start an application with the following environment variable: DOTNET_DiagnosticPorts=/home/user/myport.sock
    
  2. In a separate console, launch the target application with the environment variable DOTNET_DiagnosticPorts set to the value in the dotnet-trace output.

    export DOTNET_DiagnosticPorts=/home/user/myport.sock
    ./my-dotnet-app arg1 arg2
    

    This should then enable dotnet-trace to start tracing my-dotnet-app:

    Waiting for connection on myport.sock
    Start an application with the following environment variable: DOTNET_DiagnosticPorts=myport.sock
    Starting a counter session. Press Q to quit.
    

    Important

    Launching your app with dotnet run can be problematic because the dotnet CLI may spawn many child processes that are not your app and they can connect to dotnet-trace before your app, leaving your app to be suspended at run time. It is recommended you directly use a self-contained version of the app or use dotnet exec to launch the application.

View the trace captured from dotnet-trace

On Windows, you can view .nettrace files in Visual Studio or PerfView for analysis.

On Linux, you can view the trace by changing the output format of dotnet-trace to speedscope. Change the output file format by using the -f|--format option. You can choose between nettrace (the default option) and speedscope. The option -f speedscope will make dotnet-trace produce a speedscope file. Speedscope files can be opened at https://www.speedscope.app.

For traces collected on non-Windows platforms, you can also move the trace file to a Windows machine and view it in Visual Studio or PerfView.

Note

The .NET Core runtime generates traces in the nettrace format. The traces are converted to speedscope (if specified) after the trace is completed. Since some conversions may result in loss of data, the original nettrace file is preserved next to the converted file.

Use dotnet-trace to collect counter values over time

dotnet-trace can:

  • Use EventCounter for basic health monitoring in performance-sensitive environments. For example, in production.
  • Collect traces so they don't need to be viewed in real time.

For example, to collect runtime performance counter values, use the following command:

dotnet-trace collect --process-id <PID> --providers System.Runtime:0:1:EventCounterIntervalSec=1

The preceding command tells the runtime counters to report once every second for lightweight health monitoring. Replacing EventCounterIntervalSec=1 with a higher value (for example, 60) allows collection of a smaller trace with less granularity in the counter data.

The following command reduces overhead and trace size more than the preceding one:

dotnet-trace collect --process-id <PID> --providers System.Runtime:0:1:EventCounterIntervalSec=1,Microsoft-Windows-DotNETRuntime:0:1,Microsoft-DotNETCore-SampleProfiler:0:1

The preceding command disables runtime events and the managed stack profiler.

Use .rsp file to avoid typing long commands

You can launch dotnet-trace with an .rsp file that contains the arguments to pass. This can be useful when enabling providers that expect lengthy arguments or when using a shell environment that strips characters.

For example, the following provider can be cumbersome to type out each time you want to trace:

dotnet-trace collect --providers Microsoft-Diagnostics-DiagnosticSource:0x3:5:FilterAndPayloadSpecs="SqlClientDiagnosticListener/System.Data.SqlClient.WriteCommandBefore@Activity1Start:-Command;Command.CommandText;ConnectionId;Operation;Command.Connection.ServerVersion;Command.CommandTimeout;Command.CommandType;Command.Connection.ConnectionString;Command.Connection.Database;Command.Connection.DataSource;Command.Connection.PacketSize\r\nSqlClientDiagnosticListener/System.Data.SqlClient.WriteCommandAfter@Activity1Stop:\r\nMicrosoft.EntityFrameworkCore/Microsoft.EntityFrameworkCore.Database.Command.CommandExecuting@Activity2Start:-Command;Command.CommandText;ConnectionId;IsAsync;Command.Connection.ClientConnectionId;Command.Connection.ServerVersion;Command.CommandTimeout;Command.CommandType;Command.Connection.ConnectionString;Command.Connection.Database;Command.Connection.DataSource;Command.Connection.PacketSize\r\nMicrosoft.EntityFrameworkCore/Microsoft.EntityFrameworkCore.Database.Command.CommandExecuted@Activity2Stop:",OtherProvider,AnotherProvider

In addition, the previous example contains " as part of the argument. Because quotes are not handled equally by each shell, you may experience various issues when using different shells. For example, the command to enter in zsh is different to the command in cmd.

Instead of typing this each time, you can save the following text into a file called myprofile.rsp.

--providers
Microsoft-Diagnostics-DiagnosticSource:0x3:5:FilterAndPayloadSpecs="SqlClientDiagnosticListener/System.Data.SqlClient.WriteCommandBefore@Activity1Start:-Command;Command.CommandText;ConnectionId;Operation;Command.Connection.ServerVersion;Command.CommandTimeout;Command.CommandType;Command.Connection.ConnectionString;Command.Connection.Database;Command.Connection.DataSource;Command.Connection.PacketSize\r\nSqlClientDiagnosticListener/System.Data.SqlClient.WriteCommandAfter@Activity1Stop:\r\nMicrosoft.EntityFrameworkCore/Microsoft.EntityFrameworkCore.Database.Command.CommandExecuting@Activity2Start:-Command;Command.CommandText;ConnectionId;IsAsync;Command.Connection.ClientConnectionId;Command.Connection.ServerVersion;Command.CommandTimeout;Command.CommandType;Command.Connection.ConnectionString;Command.Connection.Database;Command.Connection.DataSource;Command.Connection.PacketSize\r\nMicrosoft.EntityFrameworkCore/Microsoft.EntityFrameworkCore.Database.Command.CommandExecuted@Activity2Stop:",OtherProvider,AnotherProvider

Once you've saved myprofile.rsp, you can launch dotnet-trace with this configuration using the following command:

dotnet-trace @myprofile.rsp

See also