Add header-checking function (#9880)

* Add basic header-checking function

* Add Slack notification logic

* Modify Netlify Functions setup

* Rework function logic

* Add missing dependencies

* Remove package-lock.json from Git

* Separate header checking function

* Isolate env check into a separate function

.gitignore

@@ -32,3 +32,8 @@ public/
 
 # User-specific editorconfig files
 .editorconfig
+
+# Netlify Functions build output
+package-lock.json
+functions/
+node_modules/

Makefile

@@ -2,6 +2,8 @@ DOCKER       = docker
 HUGO_VERSION = 0.49
 DOCKER_IMAGE = kubernetes-hugo
 DOCKER_RUN   = $(DOCKER) run --rm --interactive --tty --volume $(PWD):/src
+NODE_BIN     = node_modules/.bin
+NETLIFY_FUNC = $(NODE_BIN)/netlify-lambda
 
 .PHONY: all build sass build-preview help serve
 
@@ -16,6 +18,9 @@ build: ## Build site with production settings and put deliverables in ./public
 build-preview: ## Build site with drafts and future posts enabled
 	hugo -D -F
 
+functions-build:
+	$(NETLIFY_FUNC) build functions-src
+
 check-headers-file:
 	scripts/check-headers-file.sh
 

functions-src/deploy-succeeded.js

@@ -0,0 +1,77 @@
+"use strict";
+
+const
+  { IncomingWebhook } = require('@slack/client'),
+  kubernetesSiteRoot = 'https://kubernetes.io',
+  fetch = require('node-fetch').default,
+  { SLACK_WEBHOOK_URL } = process.env;
+
+const webhook = new IncomingWebhook(SLACK_WEBHOOK_URL);
+
+// A random smattering of Kubernetes documentation pages
+// We can add as many pages here as we'd like
+const kubernetesEndpoints = [
+  'docs/home',
+  'docs/tutorials/configuration/configure-redis-using-configmap',
+
+]
+
+// Ensure that the SLACK_WEBHOOK_URL environment variable is set
+const checkEnv = () => {
+  if (!SLACK_WEBHOOK_URL) {
+    return {
+      statusCode: 422,
+      body: "[FAILURE] The Slack webhook URL must be set via the SLACK_WEBHOOK_URL environment variable"
+    }
+  }
+}
+
+// This function posts a warning message to Slack
+const sendSlackMessage = (msg) => {
+  const slackMessageObject = {
+    username: "noindex checker",
+    text: msg
+  }
+
+  // Send the message to the webhook
+  webhook.send(slackMessageObject, (err, res) => {
+    return (err) ? { statusCode: 422, body: `[ERROR] Slack webhook error: ${err}` } :
+      { statusCode: 200, body: `[SUCCESS] Response received from Slack: ${JSON.stringify(res)}` };
+  });
+}
+
+// Iterate through each Kubernetes endpoint to check for noindex headers
+const checkEndpoints = () => {
+  kubernetesEndpoints.forEach((endpoint) => {
+    const url = `${kubernetesSiteRoot}/${endpoint}`;
+
+    fetch(url)
+      .then(res => {
+        const headers = res.headers;
+
+        if ('x-robots-tag' in headers.raw() && (headers.get('x-robots-tag') == 'noindex')) {
+          const msg = `[WARNING] "X-Robots-Tag: noindex" header found on the following page: ${url}`;
+
+          // Send Slack notification
+          sendSlackMessage(msg);
+
+          return { statusCode: 404, body: msg };
+        } else {
+          const msg = `[SUCCESS] No improper X-Robots-Tag: noindex headers found on ${url}`;
+          
+          return { statusCode: 200, body: msg };
+        }
+      })
+      .catch(err => {
+        return { statusCode: 422, body: err };
+      });
+  });
+}
+
+// The handler function
+exports.handler = async (event, context) => {
+  checkEnv();
+
+  // Below are the various deploy succeeded checks
+  checkEndpoints();
+}

netlify.toml

@@ -3,6 +3,7 @@
 # It is turned off for only for the production site by using [context.master] below
 # DO NOT REMOVE THIS (contact @chenopis or @sig-docs-maintainers)
 publish = "public"
+functions = "functions"
 command = "make non-production-build"
 
 [build.environment]

package.json

@@ -0,0 +1,10 @@
+{
+  "private": true,
+  "devDependencies": {
+    "@babel/core": "^7.0.1",
+    "@slack/client": "^4.4.0",
+    "babel-loader": "^8.0.2",
+    "netlify-lambda": "^0.4.0",
+    "node-fetch": "^2.2.0"
+  }
+}