This quickstart guides you through the basics of using
Component Inspection. You will learn how to integrate Inspect into
your component using the language-specific libraries and review the data using
ffx inspect
.
For a more detailed walkthrough of Inspect concepts, see the Inspect codelab.
Project setup
See below for the quick start guide in your language of choice:
C++
This section assumes you are writing an asynchronous component and that
some part of your component (typically main.cc
) looks like this:
async::Loop loop(&kAsyncLoopConfigAttachToCurrentThread);
auto context_ = sys::ComponentContext::CreateAndServeOutgoingDirectory();
// ...
loop.Run();
This sets up an async loop, creates a ComponentContext
wrapping handles
provided by the runtime, and then runs that loop following some other
initialization work.
Add the Inspect library dependencies to your BUILD.gn
file:
"//sdk/lib/inspect/component/cpp",
"//sdk/lib/sys/cpp",
Add the following includes:
#include <lib/inspect/component/cpp/component.h>
Add the following code to initialize Inspect:
inspector_ = std::make_unique<inspect::ComponentInspector>(async_get_default_dispatcher(),
inspect::PublishOptions{});
You are now using Inspect! Create properties in the Inspect tree by attaching them to the root node:
// Attach properties to the root node of the tree
inspect::Node& root_node = inspector_->root();
// Important: Hold references to properties and don't let them go out of scope.
auto total_requests = root_node.CreateUint("total_requests", 0);
auto bytes_processed = root_node.CreateUint("bytes_processed", 0);
See Supported Data Types for a full list of data types you can try.
Health checks
The health check subsystem provides a standardized inspection metric for component health. You can use the health node to report the overall status of your component:
inspector_->Health().StartingUp();
// ...
inspector_->Health().Ok();
Testing
To test your inspect code, you can use //sdklib/inspect/testing/cpp/inspect.h:
#include <lib/inspect/cpp/inspect.h>
#include <lib/inspect/testing/cpp/inspect.h>
#include <gtest/gtest.h>
using namespace inspect::testing;
This library includes a full set of matchers to validate the contents of the Inspect tree.
// Validate the contents of the tree match
auto hierarchy_result = inspect::ReadFromVmo(inspector_.DuplicateVmo());
ASSERT_TRUE(hierarchy_result.is_ok());
EXPECT_THAT(hierarchy_result.take_value(),
NodeMatches(AllOf(PropertyList(::testing::UnorderedElementsAre(
UintIs("bytes_processed", 24), UintIs("total_requests", 2))))));
Rust
This section assumes you are writing an asynchronous component and that some
part of your component (typically main.rs
) looks similar to this:
async fn main() -> Result<(), Error> {
// ...
let mut service_fs = ServiceFs::new();
// ...
service_fs.take_and_serve_directory_handle().unwrap();
service_fs.collect::<()>().await;
Ok(())
}
Add the Inspect library dependencies to your BUILD.gn
file:
"//src/lib/diagnostics/inspect/runtime/rust",
"//src/lib/diagnostics/inspect/rust",
Add the following code to initialize Inspect:
// This creates the root of an Inspect tree
// The Inspector is a singleton that you can access from any scope
let inspector = fuchsia_inspect::component::inspector();
// This serves the Inspect tree, converting failures into fatal errors
let _inspect_server_task =
inspect_runtime::publish(inspector, inspect_runtime::PublishOptions::default());
You are now using Inspect! Create properties in the Inspect tree by attaching them to the root node:
// Attach properties to the root node of the tree
let root_node = inspector.root();
let total_requests = root_node.create_uint("total_requests", 0);
let bytes_processed = root_node.create_uint("bytes_processed", 0);
See Supported Data Types for a full list of data types you can try.
Health checks
The health check subsystem provides a standardized inspection metric for component health. You can use the health node to report the overall status of your component:
fuchsia_inspect::component::health().set_starting_up();
// ...
fuchsia_inspect::component::health().set_ok();
Testing
To test your Inspect code, you can use assert_data_tree
to validate the
contents of the Inspect tree:
// Get a reference to the root node of the Inspect tree
let inspector = fuchsia_inspect::component::inspector();
// ...
// Validate the contents of the tree match
diagnostics_assertions::assert_data_tree!(inspector, root: {
total_requests: 2u64,
bytes_processed: 24u64,
});
Inspect Libraries
Now that you have a root_node
you may start building your
hierarchy. This section describes some important concepts and patterns
to help you get started.
- A Node may have any number of key/value pairs called Properties.
- The key for a Value is always a UTF-8 string, the value may be one of the supported types below.
A Node may have any number of children, which are also Nodes.
C++
The code above gives you access to a single node named
"root". hello_world_property
is a Property that contains a string value
(aptly called a StringProperty).
- Values and Nodes are created under a parent Node.
Class Node
has creator methods for every type of
supported value. hello_world_property
was created using
CreateStringProperty
. You could create a child under the root node
by calling root_node.CreateChild("child name")
. Note that names must
always be UTF-8 strings.
- Values and Nodes have strict ownership semantics.
hello_world_property
owns the Property. When it is destroyed (goes
out of scope) the underlying Property is deleted and no longer present
in your component's Inspect output. This is true for child Nodes as well.
If you are creating a value that doesn't need to be modified, use a
ValueList
to keep them alive until they are no longer needed.
- Inspection is best-effort.
Due to space limitations, the Inspect library may be unable to satisfy
a Create
request. This error is not surfaced to your code: you will
receive a Node/Property object for which the methods are no-ops.
- Pattern: Pass in child Nodes to child objects.
It is useful to add an inspect::Node
argument to the constructors
for your own classes. The parent object, which should own its own
inspect::Node
, may then pass in the result of CreateChild(...)
to its children when they are constructed:
class Child {
public:
Child(inspect::Node my_node) : my_node_(std::move(my_node)) {
// Create a string that doesn't change, and emplace it in the ValueList
my_node_.CreateString("version", "1.0", &values_);
// Create metrics and properties on my_node_.
}
private:
inspect::Node my_node_;
inspect::StringProperty some_property_;
inspect::ValueList values_;
// ... more properties and metrics
};
class Parent {
public:
// ...
void AddChild() {
// Note: inspect::UniqueName returns a globally unique name with the specified prefix.
children_.emplace_back(my_node_.CreateChild(inspect::UniqueName("child-")));
}
private:
std::vector<Child> children_;
inspect::Node my_node_;
};
Rust
The Rust library provides two ways of managing nodes and properties: creation and recording.
With the create_*
methods, the ownership of the property or node object belongs
to the caller. When the returned object is dropped, the property is removed.
For example:
{
let property = root.create_int("name", 1);
}
In this example, property
went out of scope so a drop on the property is
called. Readers won't see this property.
With the record_*
methods, the lifetime of the property is tied to the parent
node. When the node is deleted, the recorded property is deleted.
{
let node = root.create_child("name");
{
node.record_uint(2); // no return
}
// The uint property will still be visible to readers.
}
In this example, the uint property associated with name
is visible to readers
until the parent node
goes out of scope.
Dynamic values
This section describes support in the Inspect libraries for nodes that are inflated lazily at read-time. The methods accept a callback function instead of a value. The callback function is invoked when the property value is read.
C++
The C++ library has two property creators for dynamic values:
CreateLazyNode
and CreateLazyValues
.
Both of these methods take a callback returning a promise for an
inspect::Inspector
, the only difference is how the dynamic values are
stored in the tree.
root->CreateLazyNode(name, callback)
creates a child node of
root
with the given name
. The callback
returns a promise for an
inspect::Inspector
whose root node is spliced into the parent hierarchy
when read. The example below shows that a child called "lazy" exists with
the string property "version" and has an additional child that is called
"lazy."
root->CreateLazyValues(name, callback)
works like root->CreateLazyNode(name,
callback)
, except all properties and child nodes on the promised root node are
added directly as values
to the original root
. In the second output of this example, the internal
lazy nodes do not appear and their values are flattened into properties on
root
.
root->CreateLazy{Node,Values}("lazy", [] {
Inspector a;
a.GetRoot().CreateString("version", "1.0", &a);
a.GetRoot().CreateLazy{Node,Values}("lazy", [] {
Inspector b;
b.GetRoot().RecordInt("value", 10);
return fpromise::make_ok_promise(std::move(b));
}, &a);
return fpromise::make_ok_promise(std::move(a));
});
Output (CreateLazyNode):
root:
lazy:
version = "1.0"
lazy:
value = 10
Output (CreateLazyValues):
root:
value = 10
version = "1.0"
The return value of CreateLazy{Node,Values}
is a LazyNode
that owns
the passed callback. The callback is never called once the LazyNode
is
destroyed. If you destroy a LazyNode
concurrently with the execution of
a callback, the destroy operation is blocked until the callback returns
its promise.
If you want to dynamically expose properties on this
, you may simply
write the following:
class Employee {
public:
Employee(inspect::Node node) : node_(std::move(node)) {
calls_ = node_.CreateInt("calls", 0);
// Create a lazy node that populates values on its parent
// dynamically.
// Note: The callback will never be called after the LazyNode is
// destroyed, so it is safe to capture "this."
lazy_ = node_.CreateLazyValues("lazy", [this] {
// Create a new Inspector and put any data in it you want.
inspect::Inspector inspector;
// Keep track of the number of times this callback is executed.
// This is safe because the callback is executed without locking
// any state in the parent node.
calls_.Add(1);
// ERROR: You cannot modify the LazyNode from the callback. Doing
// so may deadlock!
// lazy_ = ...
// The value is set to the result of calling a method on "this".
inspector.GetRoot().RecordInt("performance_score",
this->CalculatePerformance());
// Callbacks return a fpromise::promise<Inspector>, so return a result
// promise containing the value we created.
// You can alternatively return a promise that is completed by
// some asynchronous task.
return fpromise::make_ok_promise(std::move(inspector));
});
}
private:
inspect::Node node_;
inspect::IntProperty calls_;
inspect::LazyNode lazy_;
};
Rust
Refer to C++ Dynamic Value Support, as similar concepts apply in Rust.
Example:
root.create_lazy_{child,values}("lazy", [] {
async move {
let inspector = Inspector::default();
inspector.root().record_string("version", "1.0");
inspector.root().record_lazy_{node,values}("lazy", || {
let inspector = Inspector::default();
inspector.root().record_int("value", 10);
// `_value`'s drop is called when the function returns, so it will be removed.
// For these situations `record_` is provided.
let _value = inspector.root().create_int("gone", 2);
Ok(inspector)
});
Ok(inspector)
}
.boxed()
});
Output (create_lazy_node):
root:
lazy:
version = "1.0"
lazy:
value = 10
Output (create_lazy_values):
root:
value = 10
version = "1.0"
String references
C++
You can use inspect::StringReference
to reduce the memory footprint
of an Inspect hierarchy that has a lot of repeated data. For instance,
using inspect::Inspector;
Inspector inspector;
for (int i = 0; i < 100; i++) {
inspector.GetRoot().CreateChild("child", &inspector);
}
Will include 100 copies of the string "child"
in your inspect
output.
Alternatively,
using inspect::Inspector;
using inspect::StringReference;
namespace {
const StringReference kChild("child");
}
Inspector inspector;
for (int i = 0; i < 100; i++) {
inspector.GetRoot().CreateChild(kChild, &inspector)
}
Will generate only one copy of "child"
which is referenced 100 times.
This saves 16 bytes for each child node, and has a cost of 32 bytes for the shared data. The net result is a savings of 1568 bytes.
This pattern is recommended anywhere a global constant key would be used.
Rust
String names are automatically de-duplicated in Rust Inspect. For example,
use fuchsia_inspect::Inspector;
let inspector = Inspector::default();
for _ in 0..100 {
inspector.root().record_child("child");
}
Will generate only 1 copy of "child"
which is referenced 100 times.
This saves 16 bytes for each child node, and has a cost of 32 bytes for the shared data. The net result is a savings of 1568 bytes.
Viewing Inspect Data
You can use the ffx inspect
command to view the Inspect data
you exported from your component.
This section assumes you have SSH access to your running Fuchsia system and
that you started running your component. We will use the name
my_component.cm
as a placeholder for the name of your component's manifest.
Read your Inspect data
The command below prints the inspect hierarchies of all components running in the system:
ffx inspect show
Using the output from ffx inspect list
, you can specify a
single component (for example, my_component.cm
) as input to
ffx inspect show
:
ffx inspect show --component my_component.cm
Specifying --component
above will return data for all instances
of your component running on the system. If you know a specific
moniker of your component (for example, core/my_component
) you
may pass that instead:
ffx inspect show core/my_component
You may specify multiple components (for example, core/font_provider
and core/my_component
):
ffx inspect show core/font_provider core/my_component
You can also specify a node and property value. To see the
list of all possible selectors, use ffx inspect selectors
:
ffx inspect selectors core/my_component
You may then specify a selector (for example, core/my_component:root
)
as input to ffx inspect show
:
ffx inspect show core/my_component:root
If you don't know the moniker for your component, you may use
--component
with a selector that applies to a matched component. Note that if
there are multiple matches, you will be asked to disambiguate (all the monikers will
be listed).
ffx inspect show --component my_component.cm
This will print out the following if you followed the suggested steps above:
root:
hello = world
Print your Inspect in a testing context
Use the JSON pretty printer to obtain a full listing. For example:
use diagnostics_assertions::JsonGetter;
...
#[fuchsia::test]
fn my_test() {
let inspect = fuchsia_inspect::component::inspector();
...
print!("{}", inspect.get_pretty_json());
}
Supported Data Types
Type | Description | Notes |
---|---|---|
IntProperty | A metric containing a signed 64-bit integer. | All Languages |
UIntProperty | A metric containing an unsigned 64-bit integer. | Not supported in Dart |
DoubleProperty | A metric containing a double floating-point number. | All Languages |
BoolProperty | A metric containing a double floating-point number. | All Languages |
{Int,Double,Uint}Array | An array of metric types, includes typed wrappers for various histograms. | Same language support as base metric type |
StringArray | An array of strings. Represented as a StringReference. | Not supported in Dart. |
StringProperty | A property with a UTF-8 string value. | All Languages |
ByteVectorProperty | A property with an arbitrary byte value. | All Languages |
Node | A node under which metrics, properties, and more nodes may be nested. | All Languages |
LazyNode | Instantiates a complete tree of Nodes dynamically. | C++, Rust |