Console logs recording

You can enable this feature from your client-side by setting captureLogcat = true in our Android SDK config.

val config = PostHogAndroidConfig(apiKey = "<ph_project_api_key>").apply {
    // Enable session recording. Requires enabling in your project settings as well.
    // Default is false.
    sessionReplay = true
    // Capture logs automatically. Default is true.
    sessionReplayConfig.captureLogcat = true
    ...
}

You can enable this feature from your client-side by setting captureLog: true in our React Native SDK config.

Note: Console log recording is only available for Android.

export const posthog = new PostHog(
  '<ph_project_api_key>',
  {
    // Enable session recording. Requires enabling in your project settings as well.
    // Default is false.
    enableSessionReplay: true,
    sessionReplayConfig: {
      // Capture logs automatically. Default is true.
      // Android only (Native Logcat only)
      captureLog: true,
      ...
    },
  },
);

As console logs can contain sensitive information, we do not capture these logs automatically. You can enable this feature from your client-side by setting captureLogs = true in our iOS SDK config.

let config = PostHogConfig(apiKey: "<ph_project_api_key>", host: "https://us.i.posthog.com")
config.sessionReplay = true
config.sessionReplayConfig.captureLogs = true 

PostHogSDK.shared.setup(config)

Important: Individual console logs are truncated at 2000 characters. The rest of the log is not sent to PostHog. When truncating a log message we add ...[Truncated] to the message.

Sanitizing or skipping logs

PostHog iOS SDK allows you to customize how log messages are processed and categorized by providing a custom logSanitizer handler.

This handler is called for each new line printed to the console. It receives the line's content and can return either a processed log entry with its severity level, or nil to skip the log entirely:

let config = PostHogConfig(apiKey: "<ph_project_api_key>", host: "https://us.i.posthog.com")

// Custom log processing and sanitization
config.sessionReplayConfig.captureLogsConfig.logSanitizer = { output in

    // Skip trace-level logs
    guard !output.contains("TRACE") else {
        return nil
    }

    // Determine log level based on message content
    let level: PostHogLogLevel = {
        if output.contains("CRITICAL") { return .error }
        if output.contains("WARN") { return .warn }
        return .info
    }()
    
    // Remove sensitive information (e.g., auth tokens)
    let sanitizedMessage = output.replacingOccurrences(
        of: #"Bearer\s+[A-Za-z0-9-._~+/]+=*"#,
        with: "Bearer [TOKEN REDACTED]",
        options: .regularExpression
    )
    
    return PostHogLogEntry(level: level, message: sanitizedMessage)
}

PostHogSDK.shared.setup(config)

Configuring minimum log level

By default, the SDK only captures error-level logs to minimize noise in session replays. You can adjust this using the minLogLevel setting:

// Capture errors only (default)
config.sessionReplayConfig.captureLogsConfig.minLogLevel = .error

// Capture warnings and errors
config.sessionReplayConfig.captureLogsConfig.minLogLevel = .warn

// Capture all logs (most verbose)
config.sessionReplayConfig.captureLogsConfig.minLogLevel = .info

Only logs with a level greater than or equal to the minLogLevel will be captured. For example, if set to .warn, both warning and error logs will be captured, but info logs will be filtered out.

Default behavior

Log level

The SDK captures raw console output directly, independent of any specific logging framework or implementation. Since the raw output doesn't inherently carry log level information, the SDK deduces severity levels by analyzing the log content. By default, the SDK automatically classifies log messages into three severity levels:

• Error (highest severity)

  • Messages containing: "error", "exception", "fail", "failed"
  • OSLog entries with type "Error" or "Fault"

• Warning (medium severity)

  • Messages containing: "warning", "warn", "caution", "deprecated"
  • OSLog entries with type "Warning"

• Info (lowest severity)

  • All other messages that don't match error or warning patterns

Filtering and processing

• Log filtering based on minLogLevel (default: .error). Only logs with equal or higher severity are captured.
• Multi-line logs are automatically split into separate entries
• OSLog metadata (e.g., subsystem, category) is automatically removed for cleaner output

The SDK captures logs from:

• Standard output (stdout)
• Standard error (stderr)
• OSLog messages
• NSLog messages