Invos

05_monitoring_and_logging

3 min read

Monitoring and Logging

Effective monitoring and logging are essential for understanding application health, debugging issues, and identifying performance bottlenecks.

Cloudflare Analytics

Cloudflare provides a suite of built-in analytics tools that give us visibility into our applications.

Pages/Workers Analytics

  • Location: In the Cloudflare dashboard, under Workers & Pages.
  • Key Metrics to Monitor:
    • Invocations: The number of times your server-side functions have been run. Spikes could indicate high traffic or a potential issue.
    • CPU Time: The amount of time the CPU is active. Look for trends or outliers in the 50th, 90th, and 95th percentiles. High CPU time can indicate inefficient code.
    • Errors: The number of unhandled exceptions in your functions. This should be as close to zero as possible.
  • Alerts:
    • Configure alerts to be notified when certain thresholds are breached (e.g., a sudden spike in errors).

Web Analytics

  • Location: In the Cloudflare dashboard, under your Site’s Analytics & Logs tab.
  • Key Metrics:
    • Requests and Page Views: To understand traffic patterns.
    • Bandwidth: To monitor data transfer.
    • Cache Hit Ratio: A high cache hit ratio is good, as it means Cloudflare is serving content from its edge network, reducing load on our origin functions.
    • Security: The number of threats blocked by the WAF.

Logging

By default, you can view real-time logs from your Cloudflare Workers and Pages functions directly in the dashboard.

  • Location: From your Pages project, go to the deployment you want to inspect, and you can view live logs.
  • console.log(): Anything you console.log() in your server-side code (e.g., Next.js API routes) will appear in these logs.

Limitations of Default Logging

  • Cloudflare’s default logging is real-time and not stored for long periods.
  • It can be difficult to search and analyze large volumes of logs.

Third-Party Logging (Log Drains)

For more robust and persistent logging, we integrate with a third-party logging provider (e.g., Datadog, Logflare, Sentry).

  • How it Works: We configure a “Log Drain” from Cloudflare to send our logs to the third-party service.
  • Setup:
    1. Set up an account with the chosen logging provider.
    2. In the Cloudflare dashboard for your Site, go to Analytics & Logs > Logpush.
    3. Create a new Logpush job, select the dataset (e.g., Workers Trace Events), and configure the destination to be your logging provider’s endpoint.
  • Benefits:
    • Long-term storage: Logs are stored for days, weeks, or months.
    • Advanced Search: Powerful query languages to search and filter logs.
    • Alerting: Create alerts based on log patterns (e.g., alert if the string “FATAL” appears more than 10 times in a minute).
    • Dashboards: Create visualizations and dashboards to monitor application health from logs.

Error Tracking

For capturing and tracking application errors in detail, we use a dedicated error tracking service like Sentry.

  • Integration: The Sentry SDK is added to our Next.js application.
  • Benefits:
    • Rich Error Details: Sentry captures the stack trace, request details, user context, and more for every error.
    • Grouping and Triaging: Sentry groups similar errors together, making it easy to see which errors are most common.
    • Alerting: Sentry can send alerts to Slack or email when a new or high-priority error occurs.

By combining Cloudflare’s built-in analytics with third-party logging and error tracking services, we can maintain a comprehensive view of our application’s health and performance.

Graph View

Backlinks