Quill

Asynchronous Low Latency C++ Logging Library

• Introduction
• Features
• Performance
• Supported Platforms And Compilers
• Basic Usage
• CMake Integration
• Documentation
• License

homebrew	vcpkg	conan
brew install quill	vcpkg install quill	quill/[>=1.2.3]

Introduction

Quill is a cross-platform low latency logging library based on C++14.

The main goals of the library are:

• Simplicity A small example code snippet should be enough to get started and use most of features.
• Performance Ultra low latency for the caller threads, no formatting on the hot-path, asynchronous only mode.
• Convenience Ease application monitoring/debugging. Latency is equal to latencies of binary loggers, but the produced log is in human readable form.

Features

• Type safe python style API with compile type checks and build-in support for logging STL types/containers by using the excellent {fmt} library.
• Blazing fast. See Benchmarks.
• Formatting is performed outside of the hot-path in a backend logging thread. For non-built-in types ostream::operator<<() is called on a copy of the object by the backend logging thread. Unsafe to copy non-trivial user defined are detected in compile time. Those types can be tagged as safe-to-copy to avoid formatting them on the hot path. See User Defined Types.
• Log levels can be completely stripped out at compile time reducing if branches.
• Custom logs formatting. Logs can be formatted based on a user specified pattern. See Formatters.
• Log statements always appear in timestamp order even if produced by different threads. This makes debugging easier in multi-threaded applications.
• Backtrace logging. Store log messages in a ring buffer and display later when a higher severity log statement occurs or on demand. See Backtrace Logging.
• guaranteed non-blocking or non-guaranteed logging. In guaranteed non-blocking logging more memory is allocated so the caller doesn't get blocked, log messages are never dropped. In non-guaranteed mode there is no heap allocation log messages can be dropped. See FAQ.
• Support for wide character logging and wide character filenames (Windows only).
• Various logging targets. See Handlers.
  • Console logging with coloured log support.
  • File Logging
  • Rotating log files
  • Time rotating log files
• Filters for filtering log messages. See Filters.
• Clean warning-free codebase even on high warning levels.

Performance

The following message is logged 2'000'000 times per thread LOG_INFO(logger, "Logging int: {}, int: {}, double: {}", i, j, d).
Results are in nanoseconds.

1 Thread

Library	50th	75th	90th	95th	99th	99.9th	Worst
PlatformLab NanoLog	13	15	16	17	18	21	93
Quill v1.3.0 Bounded Queue	14	17	20	22	27	39	211
Quill v1.3.0 Unbounded Queue	16	18	21	23	28	41	185
MS BinLog	30	33	34	34	39	92	227
Reckless	70	76	81	84	91	135	5649209
Iyengar NanoLog	72	75	79	123	173	264	34295
spdlog	522	589	656	698	783	907	1595
g3log	2705	2850	2991	3085	3279	3530	3949

4 Threads

Library	50th	75th	90th	95th	99th	99.9th	Worst
PlatformLab NanoLog	13	15	16	17	18	21	150
Quill v.1.30 Bounded Queue	13	16	19	21	27	50	214
Quill v.1.30 Unbounded Queue	14	18	21	23	31	87	1377
MS BinLog	30	33	34	35	42	90	203
Reckless	99	103	106	109	134	343	31535706
Iyengar NanoLog	72	125	159	176	219	321	11402
spdlog	538	616	700	757	912	1116	1592
g3log	2637	2792	2932	3021	3214	3485	4025

The benchmarks are done on Linux (Ubuntu/RHEL) with GCC 9.1.

Each thread is pinned on a different cpu.
Running the backend logger thread in the same CPU as the caller hot-path threads, slows down the log message processing on the backend logging thread and will cause the SPSC queue to fill faster and re-allocate.

Continuously Logging messages in a loop makes the consumer (backend logging thread) unable to follow up and the queue will have to re-allocate or block for most logging libraries expect very high throughput binary loggers like PlatformLab Nanolog. Therefore, a different approach was followed that suits more to a real time application:

1. 20 messages are logged in a loop.
2. calculate/store the average latency for those messages.
3. wait between 1-2 ms.
4. repeat for n iterations.

I run each logger benchmark four times and the above latencies are the second best result.

The benchmark code can be found here.
More benchmarks results (bench_results_*.txt) can be found here.

Verdict

If you want to use a printf API and only log primitive types, PlatformLab NanoLog is the fastest logger with the lowest latencies and high throughput when we look at 99th percentile. However :

1. Need to decompress a binary log file to read log each time.
2. Need to specify the type we are logging for each call to the logger.
3. To log any user defined type or a something like std::vector via NanoLog you would first have to convert it to a string in the hot path. Instead, Quill copies the object and covertion to string is performed by the backend thread.

Quill backend is not as high throughput as NanoLog as it doesn't log binary. In terms of latency it is almost as fast as Nanolog. Quill is much more feature rich offering custom formatting, several logger objects, human readable log files and a superior format API that also supports user-defined types.

Supported Platforms And Compilers

Quill requires a C++14 compiler. Minimum required versions of supported compilers are shown in the below table. Cygwin is not supported at the moment.

Compiler	Notes
GCC	version >= 5.0
Clang	version >= 5.0
MSVC++	version >= 14.3

Platform	Notes
Linux	Ubuntu, RHEL, Centos, Fedora
Windows	Windows 10 - version 1607, Windows Server 2016
macOS	Tested with Xcode 9.4

Basic usage

#include "quill/Quill.h"

int main()
{
  quill::enable_console_colours();
  quill::start();

  quill::Logger* logger = quill::get_logger();
  logger->set_log_level(quill::LogLevel::TraceL3);

  // enable a backtrace that will get flushed when we log CRITICAL
  logger->init_backtrace(2, quill::LogLevel::Critical);

  LOG_BACKTRACE(logger, "Backtrace log {}", 1);
  LOG_BACKTRACE(logger, "Backtrace log {}", 2);

  LOG_INFO(logger, "Welcome to Quill!");
  LOG_ERROR(logger, "An error message. error code {}", 123);
  LOG_WARNING(logger, "A warning message.");
  LOG_CRITICAL(logger, "A critical error.");
  LOG_DEBUG(logger, "Debugging foo {}", 1234);
  LOG_TRACE_L1(logger, "{:>30}", "right aligned");
  LOG_TRACE_L2(logger, "Positional arguments are {1} {0} ", "too", "supported");
  LOG_TRACE_L3(logger, "Support for floats {:03.2f}", 1.23456);
}

Output

CMake-Integration

External

Building and Installing Quill as Static Library

git clone https://github.com/odygrd/quill.git
mkdir cmake_build
cd cmake_build
make install

Note: To install in custom directory invoke cmake with -DCMAKE_INSTALL_PREFIX=/quill/install-dir/

Building and Installing Quill as Static Library With External libfmt

cmake -DCMAKE_PREFIX_PATH=/my/fmt/fmt-config.cmake-directory/ -DQUILL_FMT_EXTERNAL=ON -DCMAKE_INSTALL_PREFIX=/quill/install-dir/'

Then use the library from a CMake project, you can locate it directly with find_package()

Directory Structure

my_project/
├── CMakeLists.txt
├── main.cpp

CMakeLists.txt

# Set only if needed - quill was installed under a custom non-standard directory
set(CMAKE_PREFIX_PATH /test_quill/usr/local/)

find_package(quill REQUIRED)

# Linking your project against quill
add_executable(example main.cpp)
target_link_libraries(example PRIVATE quill::quill)

main.cpp

See basic usage

Embedded

To embed the library directly, copy the source folder to your project and call add_subdirectory() in your CMakeLists.txt file

Directory Structure

my_project/
├── quill/            (source folder)
├── CMakeLists.txt
├── main.cpp

CMakeLists.txt

cmake_minimum_required(VERSION 3.1.0)
project(my_project)

set(CMAKE_CXX_STANDARD 14) 

add_subdirectory(quill)

add_executable(my_project main.cpp)
target_link_libraries(my_project PRIVATE quill::quill)

main.cpp

See basic usage

Documentation

Advanced usage and additional documentation can be found in the wiki pages.

The examples folder is also a good source of documentation.

License

Quill is licensed under the MIT License

Quill depends on third party libraries with separate copyright notices and license terms. Your use of the source code for these subcomponents is subject to the terms and conditions of the following licenses.

• (MIT License) {fmt} (https://github.com/fmtlib/fmt/blob/master/LICENSE.rst)
• (MIT License) invoke.hpp (https://github.com/BlackMATov/invoke.hpp/blob/master/LICENSE.md)
• (MIT License) doctest (https://github.com/onqtam/doctest/blob/master/LICENSE.txt)