[−][src]Crate tracing
A scoped, structured logging and diagnostics system.
Overview
tracing is a framework for instrumenting Rust programs to collect
structured, event-based diagnostic information.
In asynchronous systems like Tokio, interpreting traditional log messages can
often be quite challenging. Since individual tasks are multiplexed on the same
thread, associated events and log lines are intermixed making it difficult to
trace the logic flow. tracing expands upon logging-style diagnostics by
allowing libraries and applications to record structured events with additional
information about temporality and causality — unlike a log message, a span
in tracing has a beginning and end time, may be entered and exited by the
flow of execution, and may exist within a nested tree of similar spans. In
addition, tracing spans are structured, with the ability to record typed
data as well as textual messages.
The tracing crate provides the APIs necessary for instrumenting libraries
and applications to emit trace data.
Core Concepts
The core of tracing's API is composed of spans, events and
subscribers. We'll cover these in turn.
Spans
To record the flow of execution through a program, tracing introduces the
concept of spans. Unlike a log line that represents a moment in
time, a span represents a period of time with a beginning and an end. When a
program begins executing in a context or performing a unit of work, it
enters that context's span, and when it stops executing in that context,
it exits the span. The span in which a thread is currently executing is
referred to as that thread's current span.
For example:
use tracing::{span, Level}; let span = span!(Level::TRACE, "my_span"); // `enter` returns a RAII guard which, when dropped, exits the span. this // indicates that we are in the span for the current lexical scope. let _enter = span.enter(); // perform some work in the context of `my_span`...
The span module's documentation provides further details on how to
use spans.
Events
An Event represents a moment in time. It signifies something that
happened while a trace was being recorded. Events are comparable to the log
records emitted by unstructured logging code, but unlike a typical log line,
an Event may occur within the context of a span.
For example:
use tracing::{event, span, Level}; // records an event outside of any span context: event!(Level::INFO, "something happened"); let span = span!(Level::INFO, "my_span"); let _guard = span.enter(); // records an event within "my_span". event!(Level::DEBUG, "something happened inside my_span");
In general, events should be used to represent points in time within a span — a request returned with a given status code, n new items were taken from a queue, and so on.
The Event struct documentation provides further details on using
events.
Subscribers
As Spans and Events occur, they are recorded or aggregated by
implementations of the Subscriber trait. Subscribers are notified
when an Event takes place and when a Span is entered or exited. These
notifications are represented by the following Subscriber trait methods:
- event, called when an- Eventtakes place,
- enter, called when execution enters a- Span,
- exit, called when execution exits a- Span
In addition, subscribers may implement the enabled function to filter
the notifications they receive based on metadata describing each Span
or Event. If a call to Subscriber::enabled returns false for a given
set of metadata, that Subscriber will not be notified about the
corresponding Span or Event. For performance reasons, if no currently
active subscribers express  interest in a given set of metadata by returning
true, then the corresponding Span or Event will never be constructed.
Usage
First, add this to your Cargo.toml:
[dependencies]
tracing = "0.1"
Compiler support: requires rustc 1.39+
Recording Spans and Events
Spans and events are recorded using macros.
Spans
The span! macro expands to a Span struct which is used to
record a span. The [Span::enter] method on that struct records that the
span has been entered, and returns a RAII guard object, which will exit
the span when dropped.
For example:
use tracing::{span, Level}; // Construct a new span named "my span" with trace log level. let span = span!(Level::TRACE, "my span"); // Enter the span, returning a guard object. let _enter = span.enter(); // Any trace events that occur before the guard is dropped will occur // within the span. // Dropping the guard will exit the span.
The #[instrument] attribute provides an easy way to
add tracing spans to functions. A function annotated with #[instrument]
will create and enter a span with that function's name every time the
function is called, with arguments to that function will be recorded as
fields using fmt::Debug.
For example:
use tracing::{Level, event, instrument}; #[instrument] pub fn my_function(my_arg: usize) { // This event will be recorded inside a span named `my_function` with the // field `my_arg`. event!(Level::INFO, "inside my_function!"); // ... }
Note: using #[instrument] on async fns requires the
tracing-futures crate as a dependency, as well.
You can find more examples showing how to use this crate here.
Events
Events are recorded using the event! macro:
use tracing::{event, Level}; event!(Level::INFO, "something has happened!");
Using the Macros
The span! and event! macros use fairly similar syntax, with some
exceptions.
Configuring Attributes
Both macros require a Level specifying the verbosity of the span or
event. Optionally, the target and parent span may be overridden. If the
target and parent span are not overridden, they will default to the
module path where the macro was invoked and the current span (as determined
by the subscriber), respectively.
For example:
span!(target: "app_spans", Level::TRACE, "my span"); event!(target: "app_events", Level::INFO, "something has happened!");
let span = span!(Level::TRACE, "my span"); event!(parent: &span, Level::INFO, "something has happened!");
The span macros also take a string literal after the level, to set the name of the span.
Recording Fields
Structured fields on spans and events are specified using the syntax
field_name = field_value. Fields are separated by commas.
// records an event with two fields: // - "answer", with the value 42 // - "question", with the value "life, the universe and everything" event!(Level::INFO, answer = 42, question = "life, the universe, and everything");
As shorthand, local variables may be used as field values without an assignment, similar to struct initializers. For example:
let user = "ferris"; span!(Level::TRACE, "login", user); // is equivalent to: span!(Level::TRACE, "login", user = user);
Field names can include dots, but should not be terminated by them:
let user = "ferris"; let email = "ferris@rust-lang.org"; span!(Level::TRACE, "login", user, user.email = email);
Since field names can include dots, fields on local structs can be used using the local variable shorthand:
let user = User { name: "ferris", email: "ferris@rust-lang.org", }; // the span will have the fields `user.name = "ferris"` and // `user.email = "ferris@rust-lang.org"`. span!(Level::TRACE, "login", user.name, user.email);
The ? sigil is shorthand that specifies a field should be recorded using
its fmt::Debug implementation:
#[derive(Debug)] struct MyStruct { field: &'static str, } let my_struct = MyStruct { field: "Hello world!" }; // `my_struct` will be recorded using its `fmt::Debug` implementation. event!(Level::TRACE, greeting = ?my_struct); // is equivalent to: event!(Level::TRACE, greeting = tracing::field::debug(&my_struct));
The % sigil operates similarly, but indicates that the value should be
recorded using its fmt::Display implementation:
// `my_struct.field` will be recorded using its `fmt::Display` implementation. event!(Level::TRACE, greeting = %my_struct.field); // is equivalent to: event!(Level::TRACE, greeting = tracing::field::display(&my_struct.field));
The % and ? sigils may also be used with local variable shorthand:
// `my_struct.field` will be recorded using its `fmt::Display` implementation. event!(Level::TRACE, %my_struct.field);
Additionally, a span may declare fields with the special value Empty,
which indicates that that the value for that field does not currently exist
but may be recorded later. For example:
use tracing::{trace_span, field}; // Create a span with two fields: `greeting`, with the value "hello world", and // `parting`, without a value. let span = trace_span!("my_span", greeting = "hello world", parting = field::Empty); // ... // Now, record a value for parting as well. span.record("parting", &"goodbye world!");
Note that a span may have up to 32 fields. The following will not compile:
let bad_span = span!( Level::TRACE, "too many fields!", a = 1, b = 2, c = 3, d = 4, e = 5, f = 6, g = 7, h = 8, i = 9, j = 10, k = 11, l = 12, m = 13, n = 14, o = 15, p = 16, q = 17, r = 18, s = 19, t = 20, u = 21, v = 22, w = 23, x = 24, y = 25, z = 26, aa = 27, bb = 28, cc = 29, dd = 30, ee = 31, ff = 32, gg = 33 );
Finally, events may also include human-readable messages, in the form of a
format string and (optional) arguments, after the event's
key-value fields. If a format string and arguments are provided,
they will implicitly create a new field named message whose value is the
provided set of format arguments.
For example:
let question = "the answer to the ultimate question of life, the universe, and everything"; let answer = 42; // records an event with the following fields: // - `question.answer` with the value 42, // - `question.tricky` with the value `true`, // - "message", with the value "the answer to the ultimate question of life, the // universe, and everything is 42." event!( Level::DEBUG, question.answer = answer, question.tricky = true, "the answer to {} is {}.", question, answer );
Specifying a formatted message in this manner does not allocate by default.
Shorthand Macros
tracing also offers a number of macros with preset verbosity levels.
The trace!, debug!, info!, warn!, and error! behave
similarly to the event! macro, but with the Level argument already
specified, while the corresponding trace_span!, debug_span!,
info_span!, warn_span!, and error_span! macros are the same,
but for the span! macro.
These are intended both as a shorthand, and for compatibility with the log
crate (see the next section).
For log Users
Users of the log crate should note that tracing exposes a set of
macros for creating Events (trace!, debug!, info!, warn!, and
error!) which may be invoked with the same syntax as the similarly-named
macros from the log crate. Often, the process of converting a project to
use tracing can begin with a simple drop-in replacement.
Let's consider the log crate's yak-shaving example:
use std::{error::Error, io}; use tracing::{debug, error, info, span, warn, Level}; // the `#[tracing::instrument]` attribute creates and enters a span // every time the instrumented function is called. The span is named after the // the function or method. Paramaters passed to the function are recorded as fields. #[tracing::instrument] pub fn shave(yak: usize) -> Result<(), Box<dyn Error + 'static>> { // this creates an event at the DEBUG level with two fields: // - `excitement`, with the key "excitement" and the value "yay!" // - `message`, with the key "message" and the value "hello! I'm gonna shave a yak." // // unlike other fields, `message`'s shorthand initialization is just the string itself. debug!(excitement = "yay!", "hello! I'm gonna shave a yak."); if yak == 3 { warn!("could not locate yak!"); // note that this is intended to demonstrate `tracing`'s features, not idiomatic // error handling! in a library or application, you should consider returning // a dedicated `YakError`. libraries like snafu or thiserror make this easy. return Err(io::Error::new(io::ErrorKind::Other, "shaving yak failed!").into()); } else { debug!("yak shaved successfully"); } Ok(()) } pub fn shave_all(yaks: usize) -> usize { // Constructs a new span named "shaving_yaks" at the TRACE level, // and a field whose key is "yaks". This is equivalent to writing: // // let span = span!(Level::TRACE, "shaving_yaks", yaks = yaks); // // local variables (`yaks`) can be used as field values // without an assignment, similar to struct initializers. let span = span!(Level::TRACE, "shaving_yaks", yaks); let _enter = span.enter(); info!("shaving yaks"); let mut yaks_shaved = 0; for yak in 1..=yaks { let res = shave(yak); debug!(yak, shaved = res.is_ok()); if let Err(ref error) = res { // Like spans, events can also use the field initialization shorthand. // In this instance, `yak` is the field being initalized. error!(yak, error = error.as_ref(), "failed to shave yak!"); } else { yaks_shaved += 1; } debug!(yaks_shaved); } yaks_shaved }
In libraries
Libraries should link only to the tracing crate, and use the provided
macros to record whatever information will be useful to downstream
consumers.
In executables
In order to record trace events, executables have to use a Subscriber
implementation compatible with tracing. A Subscriber implements a
way of collecting trace data, such as by logging it to standard output.
This library does not contain any Subscriber implementations; these are
provided by other crates.
The simplest way to use a subscriber is to call the set_global_default
function:
extern crate tracing; let my_subscriber = FooSubscriber::new(); tracing::subscriber::set_global_default(my_subscriber) .expect("setting tracing default failed");
Note: Libraries should NOT call set_global_default()! That will
cause conflicts when executables try to set the default later.
This subscriber will be used as the default in all threads for the
remainder of the duration of the program, similar to setting the logger
in the log crate.
In addition, the default subscriber can be set through using the
with_default function. This follows the tokio pattern of using
closures to represent executing code in a context that is exited at the end
of the closure. For example:
let my_subscriber = FooSubscriber::new(); tracing::subscriber::with_default(my_subscriber, || { // Any trace events generated in this closure or by functions it calls // will be collected by `my_subscriber`. })
This approach allows trace data to be collected by multiple subscribers within different contexts in the program. Note that the override only applies to the currently executing thread; other threads will not see the change from with_default.
Any trace events generated outside the context of a subscriber will not be collected.
Once a subscriber has been set, instrumentation points may be added to the
executable using the tracing crate's macros.
Related Crates
In addition to tracing and tracing-core, the tokio-rs/tracing repository
contains several additional crates designed to be used with the tracing ecosystem.
This includes a collection of Subscriber implementations, as well as utility
and adapter crates to assist in writing Subscribers and instrumenting
applications.
In particular, the following crates are likely to be of interest:
- tracing-futuresprovides a compatibility layer with the- futurescrate, allowing spans to be attached to- Futures,- Streams, and- Executors.
- tracing-subscriberprovides- Subscriberimplementations and utilities for working with- Subscribers. This includes a- FmtSubscriber- FmtSubscriberfor logging formatted trace data to stdout, with similar filtering and formatting to the- env_loggercrate.
- tracing-logprovides a compatibility layer with the- logcrate, allowing log messages to be recorded as- tracing- Events within the trace tree. This is useful when a project using- tracinghave dependencies which use- log. Note that if you're using- tracing-subscriber's- FmtSubscriber, you don't need to depend on- tracing-logdirectly.
Additionally, there are also several third-party crates which are not
maintained by the tokio project. These include:
- tracing-timingimplements inter-event timing metrics on top of- tracing. It provides a subscriber that records the time elapsed between pairs of- tracingevents and generates histograms.
- tracing-opentelemetryprovides a subscriber for emitting traces to OpenTelemetry-compatible distributed tracing systems.
- tracing-honeycombimplements a subscriber for reporting traces to honeycomb.io.
- tracing-actixprovides- tracingintegration for the- actixactor framework.
- tracing-gelfimplements a subscriber for exporting traces in Greylog GELF format.
- tracing-cozprovides integration with the coz causal profiler (Linux-only).
If you're the maintainer of a tracing ecosystem crate not listed above,
please let us know! We'd love to add your project to the list!
Note: that some of the ecosystem crates are currently unreleased and
undergoing active development. They may be less stable than tracing and
tracing-core.
Crate Feature Flags
The following crate feature flags are available:
- 
A set of features controlling the static verbosity level. 
- 
log: causes trace instrumentation points to emitlogrecords as well as trace events, if a defaulttracingsubscriber has not been set. This is intended for use in libraries whose users may be using eithertracingorlog. Note:logsupport will not work whentracingis renamed inCargo.toml, due to oddities in macro expansion.
- 
log-always: Emitlogrecords from alltracingspans and events, even atracingsubscriber has been set. This should be set only by applications which intend to collect traces and logs separately; if an adapter is used to convertlogrecords intotracingevents, this will cause duplicate events to occur.
- 
attributes: Includes support for the#[instrument]attribute. This is on by default, but does bring in thesyncrate as a dependency, which may add to the compile time of crates that do not already use it.
- 
std: Depend on the Rust standard library (enabled by default).no_stdusers may disable this feature withdefault-features = false:[dependencies] tracing = { version = "0.1.13", default-features = false }Compiler support: requires rustc 1.39+ Note: tracing'sno_stdsupport requiresliballoc.
Modules
| dispatcher | Dispatches trace events to  | 
| event | Events represent single points in time during the execution of a program. | 
| field | Structured data associated with  | 
| level_filters | Trace verbosity level filtering. | 
| span | Spans represent periods of time in which a program was executing in a particular context. | 
| subscriber | Collects and records trace data. | 
Macros
| debug | Constructs an event at the debug level. | 
| debug_span | Constructs a span at the debug level. | 
| error | Constructs an event at the error level. | 
| error_span | Constructs a span at the error level. | 
| event | Constructs a new  | 
| info | Constructs an event at the info level. | 
| info_span | Constructs a span at the info level. | 
| span | Constructs a new span. | 
| trace | Constructs an event at the trace level. | 
| trace_span | Constructs a span at the trace level. | 
| warn | Constructs an event at the warn level. | 
| warn_span | Constructs a span at the warn level. | 
Structs
| Dispatch | 
 | 
| Event | 
 | 
| Level | Describes the level of verbosity of a span or event. | 
| Metadata | |
| Span | A handle representing a span, with the capability to enter the span if it exists. | 
Traits
| Subscriber | Trait representing the functions required to collect trace data. | 
| Value | A field value of an erased type. | 
Attribute Macros
| instrument | Instruments a function to create and enter a  |