A toolkit for consumers of Node.js diagnostic Reports

• Run heuristics ("Rules") against diagnostic reports to uncover issues needing attention
  • Comes with a set of built-in Rules and a "recommended" configuration
  • Custom, extendable, and shareable Rules
• Purpose-built, "smart" diagnostic report diffing
  • "What has changed from last report?"
  • Ignores fields that don't matter
• Automatically redacts sensitive information from report output
  • Redacts cloud provider tokens, session IDs, etc., from environment variables
  • Redaction available as a single operation--use this before sending the report somewhere else!
• A friendly, colorful command-line interface
• Two public APIs: Promise-based and Observable-based

• A choice of output formats:
  • Tabular, human-readable output
  • JSON
  • CSV
• Handles one or more report files
• Written using ES modules to facilitate bundling as library for the web

🚨 WARNING! 🚨

As per semantic versioning, report-toolkit should be considered experimental until it reaches v1.0.0. Until then, the command-line options, programmatic API or output could change at any time.

report-toolkit is unlikely to reach v1.0.0 until Diagnostic Reports become a stable Node.js API.

npx report-toolkit --help

or install globally:

npm install -g report-toolkit && \
report-toolkit --help

npm install report-toolkit

and:

// my-app.js
const {inspect} = require('report-toolkit');

async function main() {
  const report = JSON.parse(process.report.getReport());
  // configuration automatically loaded from `.rtkrc.js` in CWD
  const results = await inspect(report);
  if (results.length) {
    results.forEach(result => {
      // log problem and associated rule ID
      console.log(`${result.message} (${result.id})`);
    });
  } else {
    console.log('no problems!');
  }
}

main();

Diagnostic Reports landed as an experimental feature in Node.js v11.8.0.

The Node.js documentation describes reports:

"The report is intended for development, test and production use, to capture and preserve information for problem determination. It includes JavaScript and native stack traces, heap statistics, platform information, resource usage etc. With the report option enabled, diagnostic reports can be triggered on unhandled exceptions, fatal errors and user signals, in addition to triggering programmatically through API calls."

• Please see the list of issues labeled help wanted.
• report-toolkit wants your ideas for new rules! Please create an issue.

• Documentation
  • Real website (https://ibm.github.io/report-toolkit)
  • API docs (incomplete: https://ibm.github.io/report-toolkit/api)
  • Tutorials
    • Quick start guide (https://ibm.github.io/report-toolkit/quick-start)
    • CLI usage guide (https://ibm.github.io/report-toolkit/cli)
    • Programmatic usage guide
    • Configuration guide
    • Plugin authoring guide
  • Contribution guidelines
  • CoC (https://github.com/IBM/report-toolkit/blob/master/.github/CODE_OF_CONDUCT.md)
  • Design docs & developer documentation
• "Interactive" mode -- a "wizard" to help you solve problems (this needs more fleshing out, but the idea is to help developers that may not know "where to start" with diagnostic reports)

• report-toolkit-as-a-service: send reports to a service which returns inspection or diff results
• Client-side ("in your app") wrapper for report transmission or direct invocation of report-toolkit
• Use CLI or API to trigger report generation from a running node process & interpret results; could be real-time
• Easy cloud deployment & integration
• "Connectors" for logging, tracing & observability tools
• Web or Electron-based GUI
• Adapters for frameworks
• IDE/editor integration: generate a report from running process, view results

• Node.js Inspector Manager (NiM) works with report-toolkit!

• boneskull

Copyright © 2019-2020, IBM. Licensed Apache-2.0