added app solutions

Author: JaimeSalas

06-monitoring/00-prometheus/.solutions/docker-compose.yml

@@ -0,0 +1,15 @@
+version: '3.8'
+
+services:
+
+  batch:
+    image: jaimesalas/prom-batch
+    build:
+      context: ./batch
+      dockerfile: ./Dockerfile
+
+  web:
+    image: jaimesalas/prom-web
+    build:
+      context: ./web
+      dockerfile: ./Dockerfile

06-monitoring/00-prometheus/.solutions/readme.md

@@ -0,0 +1,14 @@
+# Publish images
+
+```bash
+cd .solutions
+```
+
+```bash
+docker-compose build
+```
+
+```bash
+docker push jaimesalas/prom-batch
+docker push jaimesalas/prom-web
+```

06-monitoring/00-prometheus/.solutions/web/.dockerignore

@@ -1,3 +1,3 @@
 node_modules
-Dockerfile
-*.md
+*.md 
+Dockerfile

06-monitoring/00-prometheus/.solutions/web/custom-metrics.js

@@ -0,0 +1,23 @@
+const promClient = require('prom-client');
+
+const counter = new promClient.Counter({
+  name: 'http_requests_received_total',
+  help: 'Total number of requests received and response code',
+  labelNames: ['method', 'response_code']
+});
+
+
+const histogram = new promClient.Histogram({
+  name: 'http_request_duration_seconds',
+  help: 'Shows processing requests durations',
+  labelNames: ['code']
+});
+
+const delaySummary = new promClient.Summary({
+  name: 'web_delay_seconds',
+  help: 'Custom delay added',
+});
+
+module.exports.counter = counter;
+module.exports.histogram = histogram;
+module.exports.delaySummary = delaySummary;

06-monitoring/00-prometheus/.solutions/web/index.js

@@ -1,15 +1,44 @@
 const express = require('express');
+const promClient = require('prom-client');
 const { port } = require('./config');
-const { getQuote } = require('./helpers');
+const { getQuote, delay, getRandom } = require('./helpers');
+const { counter, histogram, delaySummary } = require('./custom-metrics');
+
+const { register, collectDefaultMetrics } = promClient;
+
+collectDefaultMetrics();
 
 const app = express();
 
-// curl -X GET "http://localhost:3000/quote"
-app.get('/quote', async (_, res) => {
+// curl http://localhost:3000/metrics
+app.get('/metrics', async (_, res) => {
   try {
+    res.set('Content-Type', register.contentType);
+    res.end(await register.metrics());
+  } catch (ex) {
+    res.status(500).end(ex);
+  }
+});
+
+
+// curl -X GET "http://localhost:3000/quote?slow=true"
+app.get('/quote', async (req, res) => {
+  try {
+    counter.labels('GET', 200).inc();
+    const end = histogram.startTimer();
     const quote = getQuote();
+    const { slow } = req.query;
+
+    if (slow === 'true') {
+      const delaySeconds = getRandom(2, 10);
+      delaySummary.observe(delaySeconds);
+      await delay(delaySeconds);
+    }
+
+    histogram.labels(200).observe(end());
     res.send(quote);
   } catch (ex) {
+    counter.labels('GET', 500).inc();
     res.status(500).end(ex);
   }
 });

06-monitoring/00-prometheus/.solutions/web/package.json

@@ -16,4 +16,4 @@
   "keywords": [],
   "author": "",
   "license": "ISC"
-}
+}

06-monitoring/00-prometheus/03-summaries-histograms/01-nodejs-client-library/readme.md

@@ -11,4 +11,146 @@ El ejemplo que vamos a crear se trata de una aplicación web que expone un endpo
 Creamos `custom-metrics.js`
 
 ```js
+const promClient = require('prom-client');
+
+const counter = new promClient.Counter({
+  name: 'http_requests_received_total',
+  help: 'Total number of requests received and response code',
+  labelNames: ['method', 'response_code']
+});
+
+
+const histogram = new promClient.Histogram({
+  name: 'http_request_duration_seconds',
+  help: 'Shows processing requests durations',
+  labelNames: ['code']
+});
+
+const delaySummary = new promClient.Summary({
+  name: 'web_delay_seconds',
+  help: 'Custom delay added',
+});
+
+module.exports.counter = counter;
+module.exports.histogram = histogram;
+module.exports.delaySummary = delaySummary;
+```
+
+## 2. Actualizamos index.js
+
+Abrimos el fichero `index.js`
+
+Primero vamos añadir las métricas que se exponen por defecto:
+
+```diff
+const express = require('express');
++const promClient = require('prom-client');
+const { port } = require('./config');
+const { getQuote } = require('./helpers');
+
++const { register, collectDefaultMetrics } = promClient;
++
++collectDefaultMetrics();
+
+const app = express();
+
+// curl -X GET "http://localhost:3000/quote"
+app.get('/quote', async (_, res) => {
+  try {
+    const quote = getQuote();
+    res.send(quote);
+  } catch (ex) {
+    res.status(500).end(ex);
+  }
+});
+
+app.listen(port, () => {
+  console.log(`Listenning at ${port}`);
+});
+```
+
+Vamos a hacer que el endpoint se comporte de una manera distinta en función de recibir un `query param`. Además queremos resgistrar cada vez que se realiza una petición al endpoint `quote`
+
+```diff
+const express = require('express');
+const promClient = require('prom-client');
+const { port } = require('./config');
+-const { getQuote } = require('./helpers');
++const { getQuote, delay, getRandom } = require('./helpers');
++const { counter, histogram, delaySummary } = require('./custom-metrics');
+# ....
+```
+
+Y actualizamos el endpoint de la siguiente manera.
+
+```js
+// curl -X GET "http://localhost:3000/quote?slow=true"
+app.get('/quote', async (req, res) => {
+  try {
+    counter.labels('GET', 200).inc();
+    const end = histogram.startTimer();
+    const quote = getQuote();
+    const { slow } = req.query;
+
+    if (slow === 'true') {
+      const delaySeconds = getRandom(2, 10);
+      delaySummary.observe(delaySeconds);
+      await delay(delaySeconds);
+    }
+
+    histogram.labels(200).observe(end());
+    res.send(quote);
+  } catch (ex) {
+    counter.labels('GET', 500).inc();
+    res.status(500).end(ex);
+  }
+});
+```
+
+Para que esto funcione debemos exponer las métricas.
+
+```js
+// curl http://localhost:3000/metrics
+app.get('/metrics', async (_, res) => {
+  try {
+    res.set('Content-Type', register.contentType);
+    res.end(await register.metrics());
+  } catch (ex) {
+    res.status(500).end(ex);
+  }
+});
+```
+
+## 3. Comprobamos la app
+
+Vamos a comprobar que la aplicación funciona como nosotros esperamos
+
+```bash
+npm start
+```
+
+En un terminal nuevo hagamos un par de peticiones:
+
+```bash
+curl -X GET "http://localhost:3000/quote?slow=true"
+curl -X GET "http://localhost:3000/quote"
+```
+
+Y podemos obtener las métricas llamando a su endpoint:
+
+```bash
+curl http://localhost:3000/metrics
+```
+
+```ini
+# HELP http_requests_received_total Total number of requests received and response code
+# TYPE http_requests_received_total counter
+http_requests_received_total{method="GET",response_code="200"} 2
+
+# HELP http_request_duration_seconds Shows processing requests durations
+# TYPE http_request_duration_seconds histogram
+http_request_duration_seconds_bucket{le="0.005"} 0
+http_request_duration_seconds_bucket{le="0.01"} 0
+http_request_duration_seconds_bucket{le="0.025"} 0
+# ....
 ```

06-monitoring/00-prometheus/03-summaries-histograms/02-usando-summaries-histograms/Vagrantfile

@@ -0,0 +1,14 @@
+$script = <<-SCRIPT
+echo Downloading prometheus node exporter linux
+
+apt-get update
+apt-get -y install prometheus-node-exporter
+
+SCRIPT
+
+Vagrant.configure("2") do |config|
+  config.vm.box = "bento/ubuntu-20.04"
+  config.vm.network "private_network", ip: "10.0.0.10"
+  config.vm.provision "docker"
+  config.vm.provision "shell", inline: $script
+end