Daggy - Data Aggregation Utility and C/C++ developer library for data streams catching

Daggy main goals are server-less, cross-platform, simplicity and ease-of-use.

Daggy can be helpful for developers, QA, DevOps and engineers for debug, analyze and control any data streams, including requests and responses, in distributed network systems, for example, based on micro-service architecture.

{% hint style="info" %} In short terms, daggy run local or remote processes at the same time, simultaneously read output from processes, stream and aggregate them under once session {% endhint %}

{% embed url="https://youtu.be/oeNSwv9oYDc" %} Daggy Screencast {% endembed %}

• Introduction and goal concepts
  • Daggy High Level Design
  • Basic terms
• Getting Started
  • Getting Daggy
    • Fedora
    • Windows
    • Linux
    • MacOS
    • Install from source with conan
    • Install from source with cmake (choose for maintainers)
    • Add as conan package dependency
  • Check installation of Daggy Core C++17/20 interface
  • Check installation of Daggy Core C11 interface
  • Check installation of Daggy Console application
• Getting Started data aggregation and streaming with Daggy Console Application
  • Simple Sources
    • Example of Data Aggregation Sources with multiple commands and remote data aggregation and streaming

Table of contents generated with markdown-toc

The Daggy Project consist of:

1. Core - library for streams aggregation and catching
2. Daggy - console application for aggregation streams into files

Daggy High Level Design

The main goal of Daggy Software System is obtaining the data from envorinments that located in sources to streams into aggregators and via providers.

Environment contains data for streams. Out of box, Core supports local and remote environments, but can be extended by user defined environments. Local Environment is located on the same host, that Daggy Core instance. Remote Environment is located on the different from Daggy Core instance host. User defined environment can be located anywhere, likes databases, network disks, etc.

Sources are declarations, how to obtain the data from environments. It descirbes, which kind of data need to be conveted to streams and which provider will need.

There is example of sources that contains once local environment and once remote environment:

aliases:  
    - &my_commands
        pingYa:
            exec: ping ya.ru
            extension: log
        pingGoo:
            exec: ping goo.gl
            extension: log
        
    - &ssh_auth
        user: {{env_USER}}
        passphrase: {{env_PASSWORD}}
            
sources:
    local_environment:
        type: local
        commands: *my_commands
    remote_environment:
        host: 192.168.1.9
        type: ssh2
        parameters: *ssh_auth
        commands: *my_commands

The streams from local environment are generates via local provider (looks at type: local).

The streams from remote environment are generates via ssh2 provider (looks at type: ssh2).

Out of box Core provides local and ssh2 providers. Both providers obtains the data for streams from processes - the local provider runs local process and generates streams from process channels (stdin and stdout). Ssh2 provider runs remote processes via ssh2 protocol and also generates streams from process channels. The Daggy Core can be extended by user defined provider that will generate streams, for example, from http environment.

Providers generate streams by parts via commands. The each part has unique seq_num value, uninterruptedly and consistently. It means, that full data from stream can be obtain by adding parts of stream in seq_num ascending order. Each stream can be generated by command.

The Core translates streams from any count of providers in once Core Streams Session. The streams from Core Streams Session can be aggregated by aggregators or viewed by user.

Out of box, the Core provides several types of aggregators:

1. File - aggregates streams into files at runtime, as data arrives. This aggregator is used by Daggy Console Application.
2. Console - aggreagates streams into console output. This aggregator is used by Daggy Console Application.
3. Callback - aggregates streams into ANSI C11 callbacks. This aggregator is used by Core ANSI C11 Interface.

The Core library can be extended by user defined aggregators.

sudo dnf install daggy daggy-devel

Download installer or portable version from releases page.

Download rpm/deb or portable version from releases page.

Download portable version from releases page or install via homebrew:

brew tap synacker/daggy
brew install --build-from-source daggy

{% hint style="info" %} Build requirenments: Conan, cmake, git and C++17/20 compiler. {% endhint %}

git clone https://github.com/synacker/daggy.git
mkdir build
cd build
conan install ../daggy --build=missing -o package_deps=True
conan build ../daggy

{% hint style="info" %} System dependencies: qt6 (Core and Network), libssh2, libyaml-cpp, kainjow-mustache {% endhint %}

git clone https://github.com/synacker/daggy.git
mkdir build
cd build
cmake -DVERSION=2.1.3 ../daggy/src -DBUILD_SHARED_LIBS=ON
cmake --build .

Get daggy from conan-center.

{% code title="conanfile.py" %}

def requirements(self):
    self.requires("daggy/2.1.2")

{% endcode %}

{% code title="test.cpp" %}

#include <DaggyCore/Core.hpp>
#include <DaggyCore/Sources.hpp>
#include <DaggyCore/aggregators/CFile.hpp>
#include <DaggyCore/aggregators/CConsole.hpp>

#include <QCoreApplication>
#include <QTimer>

namespace {
constexpr const char* json_data = R"JSON(
{
    "sources": {
        "localhost" : {
            "type": "local",
            "commands": {
                "ping1": {
                    "exec": "ping 127.0.0.1",
                    "extension": "log"
                },
                "ping2": {
                    "exec": "ping 127.0.0.1",
                    "extension": "log",
                    "restart": true
                }
            }
        }
    }
}
)JSON";
}

int main(int argc, char** argv) 
{
    QCoreApplication app(argc, argv);
    daggy::Core core(*daggy::sources::convertors::json(json_data));

    daggy::aggregators::CFile file_aggregator("test");
    daggy::aggregators::CConsole console_aggregator("test");

    core.connectAggregator(&file_aggregator);
    core.connectAggregator(&console_aggregator);

    QObject::connect(&core, &daggy::Core::stateChanged, &core,
    [&](DaggyStates state){
        if(state == DaggyFinished)
            app.quit();      
    });

    QTimer::singleShot(3000, &core, [&]()
    {
        core.stop();
    });

    QTimer::singleShot(5000, &core, [&]()
    {
        app.exit(-1);
    });

    core.prepare();
    core.start();

    return app.exec();
}

{% endcode %}

{% code title="test.c" %}

#include <stdio.h>
#ifdef _WIN32
#include <Windows.h>
#else
#include <unistd.h>
#endif

#include <DaggyCore/Core.h>

const char* json_data =
"{\
    \"sources\": {\
        \"localhost\" : {\
            \"type\": \"local\",\
            \"commands\": {\
                \"ping1\": {\
                    \"exec\": \"ping 127.0.0.1\",\
                    \"extension\": \"log\"\
                },\
                \"ping2\": {\
                    \"exec\": \"ping 127.0.0.1\",\
                    \"extension\": \"log\"\
                    }\
            }\
        }\
    }\
}"
;

void sleep_ms(int milliseconds)
{
    #ifdef WIN32
        Sleep(milliseconds);
    #elif _POSIX_C_SOURCE >= 199309L
        struct timespec ts;
        ts.tv_sec = milliseconds / 1000;
        ts.tv_nsec = (milliseconds % 1000) * 1000000;
        nanosleep(&ts, NULL);
    #else
        usleep(milliseconds * 1000);
    #endif
}

int quit_after_time(void* msec)
{
    sleep_ms(*(int*)(msec));
    libdaggy_app_stop();
    return 0;
}

void on_daggy_state_changed(DaggyCore core, DaggyStates state);

void on_provider_state_changed(DaggyCore core, const char* provider_id, DaggyProviderStates state);
void on_provider_error(DaggyCore core, const char* provider_id, DaggyError error);

void on_command_state_changed(DaggyCore core, const char* provider_id, const char* command_id, DaggyCommandStates state, int exit_code);
void on_command_stream(DaggyCore core, const char* provider_id, const char* command_id, DaggyStream stream);
void on_command_error(DaggyCore core, const char* provider_id, const char* command_id, DaggyError error);

int main(int argc, char** argv)
{
    DaggyCore core;
    libdaggy_app_create(argc, argv);
    libdaggy_core_create(json_data, Json, &core);
    libdaggy_connect_aggregator(core,
                                on_daggy_state_changed,
                                on_provider_state_changed,
                                on_provider_error,
                                on_command_state_changed,
                                on_command_stream,
                                on_command_error);
    libdaggy_core_start(core);
    int time = 5000;
    libdaggy_run_in_thread(quit_after_time, &time);
    return libdaggy_app_exec();
}

void on_daggy_state_changed(DaggyCore core, DaggyStates state)
{
    printf("Daggy state changed: %d\n", state);
}

void on_provider_state_changed(DaggyCore core, const char* provider_id, DaggyProviderStates state)
{
    printf("Provider %s state changed: %d\n", provider_id, state);
}

void on_provider_error(DaggyCore core, const char* provider_id, DaggyError error)
{
    printf("Provider %s error. Code: %d, Category: %s\n", provider_id, error.error, error.category);
}

void on_command_state_changed(DaggyCore core, const char* provider_id, const char* command_id, DaggyCommandStates state, int exit_code)
{
    printf("Command %s in provider %s state changed: %d\n", command_id, provider_id, state);
}

void on_command_stream(DaggyCore core, const char* provider_id, const char* command_id, DaggyStream stream)
{
    printf("Command %s in provider %s has stream from session %s: %li\n", command_id, provider_id, stream.session, stream.seq_num);
}

void on_command_error(DaggyCore core, const char* provider_id, const char* command_id, DaggyError error)
{
    printf("Command %s in provider %s has error. Code: %d, Category: %s\n", command_id, provider_id, error.error, error.category);
}

{% endcode %}

daggy --help
Usage: daggy [options] *.yaml|*.yml|*.json

Options:
  -o, --output <folder>       Set output folder
  -f, --format <json|yaml>    Source format
  -i, --stdin                 Read data aggregation sources from stdin
  -t, --timeout <time in ms>  Auto complete timeout
  -h, --help                  Displays help on commandline options.
  --help-all                  Displays help including Qt specific options.
  -v, --version               Displays version information.

Arguments:
  file                        data aggregation sources file

Create simple.yaml

sources:
    localhost:
        type: local
        commands:
            pingYa:
                exec: ping ya.ru
                extension: log

Run daggy

daggy simple.yaml

Check console output

23:07:23:977 | AppStat  | Start aggregation in 01-04-20_23-07-23-977_simple
23:07:23:977 | ProvStat | localhost       | New state: Started
23:07:23:977 | CommStat | localhost       | pingYa          | New state: Starting
23:07:23:977 | CommStat | localhost       | pingYa          | New state: Started

There are all commands from simple.yaml/simple.json are streams in 01-04-20_23-07-23-977_simple with output files

Tailing streams from Simple Data Source

tail -f 01-04-20_23-07-23-977_simple/*
64 bytes from ya.ru (87.250.250.242): icmp_seq=99 ttl=249 time=21.2 ms
64 bytes from ya.ru (87.250.250.242): icmp_seq=100 ttl=249 time=18.8 ms
64 bytes from ya.ru (87.250.250.242): icmp_seq=101 ttl=249 time=23.5 ms
64 bytes from ya.ru (87.250.250.242): icmp_seq=102 ttl=249 time=18.8 ms
64 bytes from ya.ru (87.250.250.242): icmp_seq=103 ttl=249 time=18.8 ms
64 bytes from ya.ru (87.250.250.242): icmp_seq=104 ttl=249 time=17.4 ms
64 bytes from ya.ru (87.250.250.242): icmp_seq=105 ttl=249 time=17.4 ms
64 bytes from ya.ru (87.250.250.242): icmp_seq=106 ttl=249 time=20.1 ms
64 bytes from ya.ru (87.250.250.242): icmp_seq=107 ttl=249 time=25.8 ms
64 bytes from ya.ru (87.250.250.242): icmp_seq=108 ttl=249 time=35.1 ms
64 bytes from ya.ru (87.250.250.242): icmp_seq=109 ttl=249 time=21.1 ms

Stop data aggregation and streaming

Type CTRL+C for stopping data aggregation and streaming. Type CTRL+C twice for hard stop application, without waiting cancelation of child local and remote processes.

23:07:23:977 | AppStat  | Start aggregation in 01-04-20_23-07-23-977_simple
23:07:23:977 | ProvStat | localhost       | New state: Started
23:07:23:977 | CommStat | localhost       | pingYa          | New state: Starting
23:07:23:977 | CommStat | localhost       | pingYa          | New state: Started
^C23:17:56:667 | ProvStat | localhost       | New state: Finishing
23:17:56:668 | CommStat | localhost       | pingYa          | New state: Finished. Exit code: 0
23:17:56:668 | ProvStat | localhost       | New state: Finished
23:17:56:668 | AppStat  | Stop aggregation in 01-04-20_23-07-23-977_simple

Investigate aggregated data

ls -l 01-04-20_23-07-23-977_simple/
-rw-r--r-- 1 muxa muxa 45574 апр  1 23:17 localhost-pingYa.log

aliases:  
    - &my_commands
        pingYa:
            exec: ping ya.ru
            extension: log
        pingGoo:
            exec: ping goo.gl
            extension: log
        
    - &ssh_auth
        user: {{env_USER}}
        passphrase: {{env_PASSWORD}}
            
sources:
    localhost:
        type: local
        commands: *my_commands
    remotehost:
        host: 192.168.1.9
        type: ssh2
        parameters: *ssh_auth
        commands: *my_commands
    remotehost2:
        host: 192.168.1.9
        type: ssh2
        parameters: *ssh_auth
        commands: *my_commands
    remotehost3:
        host: 192.168.1.9
        type: ssh2
        parameters: *ssh_auth
        commands: *my_commands