Metrics microservice

This is the metrics microservice. It keeps list of metrics.

This microservice is designed to manage various metrics characterizing the operation of a process. Each metric has the following characteristics:

• metric name
• up to 3 types of measurements (in string format)
• date and time is a numerical value characterizing the metric

When adding or updating a metric, statistics on the metric are automatically calculated for different time horizons (you can specify the depth of the horizon) with the calculation of the average, maximum, minimum and accumulated values ​​within each of them.

Data access is provided through a set of API functions

The microservice currently supports the following deployment options:

• Deployment platforms: Standalone Process
• External APIs: HTTP/REST
• Persistence: Memory, Flat Files, MongoDB

This microservice has no dependencies on other microservices. Quick Links:

• Download Links
• Development Guide
• Deployment Guide
• Configuration Guide
• Client SDKs
  • Node.js SDK
• Communication Protocols
  • HTTP Version 1

Contract

Logical contract of the microservice is presented below. For physical implementation (HTTP/REST, GRPC, Lambda, etc.), please, refer to documentation of the specific protocol.

// Create or update metric struct
class MetricUpdateV1 {
    public name: string;
    public year: number;
    public month: number;
    public day: number;
    public hour: number;
    public minute?: number;
    public dimension1?: string;
    public dimension2?: string;
    public dimension3?: string;
    public value: number;
}
// Metric definition struct
class MetricDefinitionV1 {
    public name: string;
    public dimension1: string[];
    public dimension2: string[];
    public dimension3: string[];
}
// Metric value struct
class MetricValueSetV1 {
    public name: string;
    public time_horizon: number;
    public dimension1: string;
    public dimension2: string;
    public dimension3: string;
    public values: MetricValueV1[];
}
// Values of metric
class MetricValueV1 {
    public year?: number;
    public month?: number;
    public day?: number;
    public hour?: number;
    public minute?: number;
    public count: number;
    public sum: number;
    public max: number;
    public min: number;
}
// Time horizons
class TimeHorizonV1 {
    public static Total: number = 0;
    public static Year: number = 1;
    public static Month: number = 2;
    public static Day: number = 3;
    public static Hour: number = 4;
    public static Minute: number = 5;
}

interface IMetricsController {
    getMetricDefinitions(correlationId: string): Promise<MetricDefinitionV1[]>;
    getMetricDefinitionByName(correlationId: string, name: string): Promise<MetricDefinitionV1>;
    getMetricsByFilter(correlationId: string, filter: FilterParams, paging: PagingParams): Promise<DataPage<MetricValueSetV1>>;
    updateMetric(correlationId: string, update: MetricUpdateV1, maxTimeHorizon: number): Promise<void>;
    updateMetrics(correlationId: string, updates: MetricUpdateV1[], maxTimeHorizon: number): Promise<void>;
}

Download

Right now the only way to get the microservice is to check it out directly from github repository

git clone [email protected]:pip-services-infrastructure2/service-metrics-node.git

Pip.Service team is working to implement packaging and make stable releases available for your as zip downloadable archieves.

Run

Add config.yaml file to the root of the microservice folder and set configuration parameters. As the starting point you can use example configuration from config.example.yaml file.

Example of microservice configuration

{    
---
- descriptor: "pip-services-commons:logger:console:default:1.0"
  level: "trace"

- descriptor: "service-metrics:persistence:file:default:1.0"
  path: "./data/blobs"

- descriptor: "service-metrics:controller:default:default:1.0"

- descriptor: "service-metrics:service:commandable-http:default:1.0"
  connection:
    protocol: "http"
    host: "0.0.0.0"
    port: 3000
}

For more information on the microservice configuration see Configuration Guide.

Start the microservice using the command:

node run

Use

Inside your code get the reference to the client SDK

 import { MetricsHttpClientV1 } from 'client-metrics-node';

Define client configuration parameters.

// Client configuration
let httpConfig = ConfigParams.fromTuples(
            'connection.protocol', 'http',
            'connection.port', 3000,
            'connection.host', 'localhost'
        );
client.configure(httpConfig);

Instantiate the client and open connection to the microservice

// Create the client instance
client = new MetricssHttpClientV1();

// Connect to the microservice
try {
    await client.open(null);

    // Work with the microservice
    ...
} catch(err) {
    console.error('Connection to the microservice failed');
    console.error(err);
}

Now the client is ready to perform operations:

Update if exist metric or create otherwise:

try {
    await client.updateMetric(
        null,
        <MetricUpdateV1> {
            name: "metric1",
            dimension1: "A",
            dimension2: "B",
            dimension3: null,
            year: 2018,
            month: 8,
            day: 26,
            hour: 12,
            value: 123
        },
        TimeHorizonV1.Hour
    );
} catch(err) {
    console.error('Update/create metric are failed');
    console.error(err);
}
    

Update if exist metrics or create otherwise::

try {
    await client.updateMetrics(
        null,
        [
            <MetricUpdateV1> {
                name: "metric1",
                dimension1: "A",
                dimension2: "B",
                dimension3: null,
                year: 2018,
                month: 8,
                day: 26,
                hour: 13,
                value: 321
            },
            <MetricUpdateV1> {
                name: "metric2",
                dimension1: "A",
                dimension2: null,
                dimension3: "C",
                year: 2018,
                month: 8,
                day: 26,
                hour: 13,
                value: 321
            }        
        ],
        TimeHorizonV1.Hour  
    );
} catch(err) {
    console.error('Update/create metric are failed');
    console.error(err);
}
    

Get metrics by filter:

try {
    let page = await client.getMetricsByFilter(null,
        FilterParams.fromTuples("name", "metric1"),
        new PagingParams()
    );
    console.log("Metrics:");
    console.log(page.data);
} catch(err) {
    console.error("Can\'t get metrics by filter");
    console.error(err);
}

    

Get all metrics definitions:

try {
    let definitions = await client.getMetricDefinitions(null);
    console.log("All metrics definition:");
    console.log(definitions);
} catch(err) {
    console.error("Can\'t get metrics definitions");
    console.error(err);
}
    

Get metric definition by name:

try {
    let definition = await client.getMetricDefinitionByName(null, "metric2");
    console.log("Metric definition name %s:", definition.name);
    console.log(definition);
} catch(err) {
    console.error("Can\'t get metrics definition by name");
    console.error(err);
} 
              
    

Acknowledgements

This client SDK was created and currently maintained by Sergey Seroukhov and Levichev Dmitry.