Microsoft.Diagnostics.NETCore.Client
This is a top-level class that contains methods to send various diagnostics command to the runtime.
Wait for an available diagnostic endpoint to the runtime instance.
The amount of time to wait before cancelling the wait for the connection.
Wait for an available diagnostic endpoint to the runtime instance.
The token to monitor for cancellation requests.
A task the completes when a diagnostic endpoint to the runtime instance becomes available.
Start tracing the application and return an EventPipeSession object
An IEnumerable containing the list of Providers to turn on.
If true, request rundown events from the runtime
The size of the runtime's buffer for collecting events in MB
An EventPipeSession object representing the EventPipe session that just started.
Start tracing the application and return an EventPipeSession object
An EventPipeProvider to turn on.
If true, request rundown events from the runtime
The size of the runtime's buffer for collecting events in MB
An EventPipeSession object representing the EventPipe session that just started.
Start tracing the application and return an EventPipeSession object
An IEnumerable containing the list of Providers to turn on.
If true, request rundown events from the runtime
The size of the runtime's buffer for collecting events in MB
The token to monitor for cancellation requests.
An EventPipeSession object representing the EventPipe session that just started.
Start tracing the application and return an EventPipeSession object
An EventPipeProvider to turn on.
If true, request rundown events from the runtime
The size of the runtime's buffer for collecting events in MB
The token to monitor for cancellation requests.
An EventPipeSession object representing the EventPipe session that just started.
Start tracing the application and return an EventPipeSession object
Configuration of this EventPipeSession
The token to monitor for cancellation requests.
An EventPipeSession object representing the EventPipe session that just started.
Trigger a core dump generation.
Type of the dump to be generated
Full path to the dump to be generated. By default it is /tmp/coredump.{pid}
When set to true, display the dump generation debug log to the console.
Trigger a core dump generation.
Type of the dump to be generated
Full path to the dump to be generated. By default it is /tmp/coredump.{pid}
logging and crash report flags. On runtimes less than 6.0, only LoggingEnabled is supported.
Trigger a core dump generation.
Type of the dump to be generated
Full path to the dump to be generated. By default it is /tmp/coredump.{pid}
When set to true, display the dump generation debug log to the console.
The token to monitor for cancellation requests.
Trigger a core dump generation.
Type of the dump to be generated
Full path to the dump to be generated. By default it is /tmp/coredump.{pid}
logging and crash report flags. On runtimes less than 6.0, only LoggingEnabled is supported.
The token to monitor for cancellation requests.
Attach a profiler.
Timeout for attaching the profiler
Guid for the profiler to be attached
Path to the profiler to be attached
Additional data to be passed to the profiler
Set a profiler as the startup profiler. It is only valid to issue this command
while the runtime is paused at startup.
Guid for the profiler to be attached
Path to the profiler to be attached
Tell the runtime to resume execution after being paused at startup.
Set an environment variable in the target process.
The name of the environment variable to set.
The value of the environment variable to set.
Gets all environement variables and their values from the target process.
A dictionary containing all of the environment variables defined in the target process.
Get all the active processes that can be attached to.
IEnumerable of all the active process IDs.
Stops the given session
Creates a new configuration object for the EventPipeSession.
For details, see the documentation of each property of this object.
An IEnumerable containing the list of Providers to turn on.
The size of the runtime's buffer for collecting events in MB
If true, request rundown events from the runtime.
If true, record a stacktrace for every emitted event.
If true, request rundown events from the runtime.
- Rundown events are needed to correctly decode the stacktrace information for dynamically generated methods.
- Rundown happens at the end of the session. It increases the time needed to finish the session and, for large applications, may have important impact on the final trace file size.
- Consider to set this parameter to false if you don't need stacktrace information or if you're analyzing events on the fly.
The size of the runtime's buffer for collecting events in MB.
If the buffer size is too small to accommodate all in-flight events some events may be lost.
If true, record a stacktrace for every emitted event.
- The support of this parameter only comes with NET 9. Before, the stackwalk is always enabled and if this property is set to false the connection attempt will fail.
- Disabling the stackwalk makes event collection overhead considerably less
- Note that some events may choose to omit the stacktrace regardless of this parameter, specifically the events emitted from the native runtime code.
- If the stacktrace collection is disabled application-wide (using the env variable DOTNET_EventPipeEnableStackwalk) this parameter is ignored.
Providers to enable for this session.
==ADVERTISE PROTOCOL==
Before standard IPC Protocol communication can occur on a client-mode connection
the runtime must advertise itself over the connection. ALL SUBSEQUENT COMMUNICATION
IS STANDARD DIAGNOSTICS IPC PROTOCOL COMMUNICATION.
The flow for Advertise is a one-way burst of 34 bytes consisting of
8 bytes - "ADVR_V1\0" (ASCII chars + null byte)
16 bytes - CLR Instance Cookie (little-endian)
8 bytes - PID (little-endian)
2 bytes - future
Sends a single DiagnosticsIpc Message to the dotnet process associated with the .
An endpoint that provides a diagnostics connection to a runtime instance.
The DiagnosticsIpc Message to be sent
An that is the response message.
Sends a single DiagnosticsIpc Message to the dotnet process associated with the .
An endpoint that provides a diagnostics connection to a runtime instance.
The DiagnosticsIpc Message to be sent
An containing the response message and continuation stream.
Sends a single DiagnosticsIpc Message to the dotnet process associated with the .
An endpoint that provides a diagnostics connection to a runtime instance.
The DiagnosticsIpc Message to be sent
The token to monitor for cancellation requests.
An that is the response message.
Sends a single DiagnosticsIpc Message to the dotnet process associated with the .
An endpoint that provides a diagnostics connection to a runtime instance.
The DiagnosticsIpc Message to be sent
The token to monitor for cancellation requests.
An containing the response message and continuation stream.
Errors (HRESULT) returned for DiagnosticsServerCommandId.Error responses.
Different diagnostic message types that are handled by the runtime.
Initiates core dump generation
Starts an EventPipe session that writes events to a file when the session is stopped or the application exits.
Stops an EventPipe session.
Starts an EventPipe session that sends events out-of-proc through IPC.
Attaches a profiler to an existing process
Message header used to send commands to the .NET Core runtime through IPC.
Request type.
Remote process Id.
Connects to the underlying IPC transport and opens a read/write-able Stream
The amount of time to block attempting to connect
A stream used for writing and reading data to and from the target .NET process
Connects to the underlying IPC transport and opens a read/write-able Stream
The token to monitor for cancellation requests.
A task that completes with a stream used for writing and reading data to and from the target .NET process.
Wait for an available diagnostic endpoint to the runtime instance.
The amount of time to wait before cancelling the wait for the connection.
Wait for an available diagnostic endpoint to the runtime instance.
The token to monitor for cancellation requests.
A task that completes when a diagnostic endpoint to the runtime instance becomes available.
This will block until the diagnostic stream is provided. This block can happen if
the stream is acquired previously and the runtime instance has not yet reconnected
to the reversed diagnostics server.
Creates a reference to a .NET process's IPC Transport
using the default rules for a given pid
The pid of the target process
A reference to the IPC Transport
==ProcessInfo==
The response payload to issuing the GetProcessInfo command.
8 bytes - PID (little-endian)
16 bytes - CLR Runtime Instance Cookie (little-endian)
# bytes - Command line string length and data
# bytes - Operating system string length and data
# bytes - Process architecture string length and data
==ProcessInfo2==
The response payload to issuing the GetProcessInfo2 command.
8 bytes - PID (little-endian)
16 bytes - CLR Runtime Instance Cookie (little-endian)
# bytes - Command line string length and data
# bytes - Operating system string length and data
# bytes - Process architecture string length and data
# bytes - Managed entrypoint assembly name
# bytes - CLR product version string (may include prerelease labels)
The "string length and data" fields are variable length:
4 bytes - Length of string data in UTF-16 characters
(2 * length) bytes - The data of the string encoded using Unicode
(includes null terminating character)
Parses a ProcessInfo payload.
Parses a ProcessInfo2 payload.
Parses a ProcessInfo3 payload.
Base class representing a Diagnostics Server router factory.
This is a common base class for network-based server endpoints used when building router instances.
We have two subclases: for normal TCP/IP sockets, and another for WebSocket connections.
Subclasses should return a human and machine readable address of the server.
For TCP this should be something that can be passed as an address in DOTNET_DiagnosticPorts, for WebSocket it could be a URI.
Subclasses should return a human readable description of the server connection type ("tcp", "WebSocket", etc)
This class represent a TCP/IP server endpoint used when building up router instances.
This class represent a WebSocket server endpoint used when building up router instances.
This class represent a TCP/IP client endpoint used when building up router instances.
This class represent a IPC server endpoint used when building up router instances.
This class represent a IPC client endpoint used when building up router instances.
This class creates IPC Server - TCP Server router instances.
Supports NamedPipes/UnixDomainSocket server and TCP/IP server.
This class creates IPC Server - TCP Client router instances.
Supports NamedPipes/UnixDomainSocket server and TCP/IP client.
This class creates IPC Client - TCP Server router instances.
Supports NamedPipes/UnixDomainSocket client and TCP/IP server.
This class creates IPC Client - TCP Client router instances.
Supports NamedPipes/UnixDomainSocket client and TCP/IP client.
Class used to run different flavours of Diagnostics Server routers.
A collection of objects that allows for observability and mutability using handlers.
Accepts the first item it encounters and requests that the item is removed from the collection.
Returns an enumerator that iterates through the underlying collection.
The returned enumerator is over a copy of the underlying collection so that there are no concurrency issues.
Returns an enumerator that iterates through the underlying collection.
The returned enumerator is over a copy of the underlying collection so that there are no concurrency issues.
Disposes the by clearing all items and handlers.
All pending handlers with throw .
Adds an item so that it may be observed by a handler.
Item to add to the collection.
The item may be immediately consumed if a handler removes the item, thus it may
not be stored into the underlying list.
Returns the first item offered to the handler
or waits for a future item if no item is immediately available.
The amount of time to wait before cancelling the handler.
The first item offered to the handler.
Returns the item on which the handler completes or waits for future items
if the handler does not immediately complete.
The handler that determines on which item to complete.
The amount of time to wait before cancelling the handler.
The item on which the handler completes.
Returns the first item offered to the handler
or waits for a future item if no item is immediately available.
The token to monitor for cancellation requests.
A task that completes with the first item offered to the handler.
Returns the item on which the handler completes and waits for future items
if the handler does not immediately complete.
The handler that determines on which item to complete.
The token to monitor for cancellation requests.
A task that completes with the item on which the handler completes.
Clears all items from the collection.
Represents a runtine instance connection to a reversed diagnostics server.
An endpoint used to retrieve diagnostic information from the associated runtime instance.
The identifier of the process that is unique within its process namespace.
The unique identifier of the runtime instance.
Establishes server endpoint for runtime instances to connect when
configured to provide diagnostic endpoints in reverse mode.
Constructs the instance with an endpoint bound
to the location specified by .
The server endpoint.
On Windows, this can be a full pipe path or the name without the "\\.\pipe\" prefix.
On all other systems, this must be the full file path of the socket.
Constructs the instance with an endpoint bound
to the location specified by .
The server endpoint.
On Windows, this can be a full pipe path or the name without the "\\.\pipe\" prefix.
On all other systems, this must be the full file path of the socket.
When TcpIp is enabled, this can also be host:port of the listening socket.
If kind is WebSocket, start a Kestrel web server.
Otherwise if kind is TcpIp as a supported protocol for ReversedDiagnosticServer. When Kind is Tcp, address will
be analyzed and if on format host:port, ReversedDiagnosticServer will try to bind
a TcpIp listener to host and port, otherwise it will use a Unix domain socket or a Windows named pipe.
Starts listening at the address for new connections.
Starts listening at the address for new connections.
The maximum number of connections the server will support.
Gets endpoint information when a new runtime instance connects to the server.
The amount of time to wait before cancelling the accept operation.
An value that contains information about the new runtime instance connection.
Gets endpoint information when a new runtime instance connects to the server.
The token to monitor for cancellation requests.
A task that completes with a value that contains information about the new runtime instance connection.
Removes endpoint information from the server so that it is no longer tracked.
The runtime instance cookie that corresponds to the endpoint to be removed.
True if the endpoint existed and was removed; otherwise false.
Accept connections from the transport.
The server transport from which connections are accepted.
The token to monitor for cancellation requests.
A task that completes when the server is no longer listening at the address.