Logging and Monitoring in Mobile Development: Essence, Tools, and How to Set Up

Logging and Monitoring on iOS: os_log and CocoaLumberjack

Mobile application logging on iOS is built on os_log — Apple's built-in system, available since iOS 10. os_log writes logs to a single system log (on-device or via Console.app), supports categories, levels, and privacy tags. The combination of logging and monitoring on iOS via Console.app and xcrun allows tracking errors and performance in real time.

Log Level on iOS

os_log has levels: Default, Info, Debug, Error, Fault. In production, Error and Fault are kept — Debug and Info are disabled. Choosing the right level is part of the mobile application monitoring strategy on iOS. Privacy: %{public}@ for unprotected data, %{private}@ for personal data.

CocoaLumberjack

CocoaLumberjack is a popular library for iOS logging. Unlike os_log, it supports custom formats, asynchronous writing, and log rotation. DDLog, DDTTYLogger (console), DDFileLogger (file with rotation). For remote monitoring, CocoaLumberjack can be combined with os_log — write to a file for remote logging and to os_log for Console.app.

Logging and Monitoring on Android: Timber and Log Level

For mobile application logging on Android, Timber is used — a library by Jake Wharton that has become the standard for the platform. Timber is a wrapper over Android Log (Logcat) that automatically adds the class tag and configures Tree. Timber is part of the Android mobile application monitoring system: DebugTree for debugging, CrashlyticsTree for sending data.

Timber in Detail

Timber uses Tree — a component that decides what to do with the log. In debug builds, DebugTree (output to Logcat), in release — CrashlyticsTree (crash sending). Timber.wtf() logs fatal errors. Timber.tag("CustomTag") overrides the tag. For effective monitoring, Timber is combined with APM tools.

Log Level on Android

Android Log has levels: VERBOSE, DEBUG, INFO, WARN, ERROR, ASSERT. ProGuard/R8 removes Log.d and Log.v in release — Timber allows logging via Plant. The level choice affects the quality of mobile application monitoring — in production, WARN and ERROR are kept.

When choosing a logging tool, consider: do you need remote logging, how often logs change, and who will analyze them. For startups, os_log or Timber is enough. For enterprise — CocoaLumberjack with server sending. At IT Sectr, we use the Timber + Crashlytics bundle for Android and os_log + CocoaLumberjack for iOS.

Structured Logging

Structured logging — writing logs in JSON format instead of plain text. Example: instead of "User login failed" we write {"event": "login_failed", "user_id": 123, "reason": "invalid_password"}. This is the foundation for automatic monitoring — the ELK stack and Grafana filter, aggregate, and analyze such data.

Benefits of Structured Logging

Filtering by fields: find all logs with event=crash in the last hour. Aggregation: build a graph of error counts by app version. Integration with ELK (Elasticsearch, Logstash, Kibana) or Grafana. Structured logging is not mandatory for first releases, but becomes critical when DAU exceeds 10,000.

Remote Logging and Log Rotation

Remote logging — sending logs from the device to a server for centralized monitoring. Implemented via custom Tree (Android) or HTTP requests. Send only Error and Warning — do not waste user traffic. Log rotation: DDFileLogger (iOS) every 1 MB or 1 day. On Android, the Logcat buffer is limited.

APM and Mobile Application Monitoring: Sentry, New Relic, Datadog

Application Performance Monitoring (APM) — a class of tools for mobile application monitoring. APM measures: startup time, network requests, FPS, memory usage, crash rate. Main players: Sentry (Performance), New Relic Mobile, Datadog Mobile.

Sentry Performance

Sentry — not only crash reporting, but also performance monitoring. Sentry Performance creates distributed traces that show how long each stage took: from button press to server response. Supports iOS, Android, Flutter, React Native. Free up to 5k events per month.

New Relic and Datadog

New Relic Mobile provides metrics: session count, time in app, crash rate, network requests. Datadog Mobile is a more modern tool with real-time dashboards and alerts. Both are paid, but have free tiers. IT Sectr recommends starting with Firebase Performance and switching to paid APM when DAU exceeds 100,000.

Firebase Performance

Firebase Performance — a free mobile application monitoring tool from Google. Automatically collects metrics: app startup time, HTTP requests (URL, method, response code), FPS. For custom tracing, use the Trace API. Integration: add the SDK to build.gradle (Android) or Podfile (iOS).

Trace and Metrics

Trace — a time segment with start and end that you measure. Example: trace = FirebasePerformance.getInstance().newTrace("checkout_process"). Metrics — numeric indicators (response size, item count). Firebase Performance does not require server infrastructure — all data goes through Firebase SDK. When monitoring mobile applications, the Trace API provides context for performance analysis.