Skip to main content

Logs

Logs provide contextual information about your application's state and events that occur during a session. The Embrace SDK's logging capabilities help you track, troubleshoot, and receive alerts about important events.

What are Logs?​

In the Embrace SDK, logs are records of events with associated severity levels, timestamps, and optional attributes. They provide critical information that can help you understand:

  • Error conditions
  • Warning situations
  • Important state changes
  • Significant user actions
  • System events

Logs are especially valuable for capturing information that requires immediate attention or for troubleshooting specific issues.

When to Use Logs​

While Embrace automatically captures sessions and their associated data, there are scenarios where you may want to add custom logs:

  • Hunting difficult bugs
  • Troubleshooting for high-value users
  • Monitoring new version rollouts
  • Tracking critical business events
  • Capturing error conditions that need immediate attention

Log Limits​

Embrace enforces certain limits on logs:

  • Maximum number of logs per session: 500
  • Maximum size of logs: 10,000 bytes

Using the Log API​

See our guide on instrumentating Logs for examples on how to instrument your application.

Logs vs Other Concepts​

  • Logs vs Traces: While traces focus on performance and operation flow, logs provide contextual information about events and states
  • Logs vs Sessions: Logs are events within a session that provide additional context
  • Logs vs Breadcrumbs: Logs are more detailed, are grouped based on message, and can have alerts built on top of their aggregation. Breadcrumbs are simpler standalone markers of user journey steps

Best Practices​

  • Be selective: Excessive logging can impact performance
  • Use appropriate severity levels: Reserve error and warning levels for actual problematic situations
  • Add meaningful attributes: This helps with filtering and understanding context
  • Keep messages clear and concise: This makes troubleshooting easier
  • Consider timing: Use the timestamp parameter to accurately reflect when events actually occurred