Merge pull request #411 from GSA/lc/update-puppeteer

Update Puppeteer to v23.8.0, Configure Cache Directory, and Enhance Consumer Instances

.gitignore

@@ -45,3 +45,6 @@ venv/
 
 # data for security scan
 /security-data
+
+# Puppeteer cache
+.cache/

.puppeteerrc.cjs

@@ -0,0 +1,9 @@
+const {join} = require('path');
+
+/**
+ * @type {import("puppeteer").Configuration}
+ */
+module.exports = {
+  // Changes the cache location for Puppeteer.
+  cacheDirectory: join(__dirname, '.cache', 'puppeteer'),
+};

libs/browser/src/puppeteer-pool.ts

@@ -70,13 +70,17 @@ export const createPuppeteerPool = (
   // Create a generic pool instance.
   const pool = genericPool.createPool<puppeteer.Browser>(
     {
-      create: () => {
+      create: async () => {
         logger.info({ msg: 'Creating browser...' });
-        return puppeteer.launch(options.puppeteerArgs).then((browser) => {
+        try {
+          const browser = await puppeteer.launch(options.puppeteerArgs);
           useCounts.set(browser, 0);
           validators.set(browser, createValidator(browser, logger));
           return browser;
-        });
+        } catch (error) {
+          logger.error({ error }, `Error creating browser: ${error.message}`);
+          throw error;
+        }
       },
       destroy: async (browser) => {
         logger.info({ msg: 'Destroying browser...' });

libs/core-scanner/src/core-scanner.service.ts

@@ -45,11 +45,14 @@ export class CoreScannerService
       scanLogger.warn(`Scan filtering enabled! Only '${input.scan}' will be executed. This should never be used in production!`);
     }
 
+    scanLogger.info('Calling browser service to start scan...');
     return await this.browserService.useBrowser(async (browser) => {
       const result = this.initResultObject(input);
 
       // Iterate over the Page enum and run the scan for each page vis this.runPage()
+      scanLogger.info('Running scans for all pages...');
       for (const page in Page) {
+        scanLogger.info(`Running scan for page '${Page[page]}'...`);
         if (Object.prototype.hasOwnProperty.call(Page, page)) {
           const pageName = Page[page];
           result[pageName] = await this.runPage(pageName, input, scanLogger, browser);

vars-prod.yml

@@ -11,6 +11,6 @@ producer_disk: 5G
 scan_schedule: '0 0 * * *'
 snapshot_schedule: '0 8 * * Sun'
 
-consumer_instances: 12
-consumer_memory: 650MB
+consumer_instances: 10
+consumer_memory: 1GB
 consumer_disk: 5G