Telemetry, Like We Dreamed About
Instead of getting better at analyzing telemetry to get answers, Lightrun’s approach is enabling developers to ask better questions.
This is a new approach, unlike what we’re used to as an industry.
See how Lightrun’s Developer Observability Platform stacks up in comparison to traditional tools:
Players: Traditional Logging Platforms – Splunk, Coralogix, Logz.io, Elastic, etc…
Lightrun Logs can be added from the IDE in real time and on demand – saving on MTTR and significantly increasing Developer Productivity. In addition, Lightrun Logs are ephemeral and do not persist in the source code – significantly reducing logging & observability costs.
- Comparison Parameter Lightrun’s Approach Traditional Logging
- Speed of adding new logs Add logs on-the-fly, without redeploying or restarting the application Adding new logs requires going through a full release cycle, redeploying & restarting the application
- Persistence of logs in codebase Logs only exist at runtime, never in source code Logs exist and persist in teh codebase forever
- Effect on Logging Cost Logs are emitted and stored only during their lifetime, costing up to 60% less (read our study on the topic) Logs are emitted, persisted and charged for constantly
- Effect on MTTR (Mean Time To Resolve) Quick, speedy incident resolution process with real-time, on-demand logs Slower, less ergonomic incident resolution process that relies on writing new code to add new logs
Players: Datadog, New Relic, Dynatrace, AppDynamics, LogicMonitor, etc.
Lightrun is a best-in-class tool for real-time, on-demand live application investigation: focusing on getting the answers, instead of analyzing endless existing telemetry in order to find the needle in the haystack.
- Comparison Parameter Lightrun’s Approach Application Performance Monitoring
- User Experience Augment static information with dynamic, real-time, on-demand telemetry Analyze static telemetry only
- Type of Telemetry Granular, dynamic, developer-focused information Aggregated, static, operator-focused information
- IDE Support Fully-featured, rich IDE plugins for creating and consuming telemetry None or limited IDE integration
- Effect on MTTR (Mean Time To Resolve) Troubleshoot on-the-fly while facing unknown unknowns head on Dive through endless telemetry to find the needle in a haystack
- Cost of Operating Lightrun Actions are ephemeral and do not persist in the source code – significantly reducing logging & observability costs APM vendors charge according to the - often vary large - volume of static telemetry that they analyze, resulting in substantial APM bills.
Players: Jaeger, Zipkin, Lightstep, Honeycomb, Epsagon, etc.
Lightrun does not use hand-crafted spans and auto-tracing, instead opting to allow developers to choose what to instrument on-the-fly – enabling a real-time but granular process of investigation.
- Comparison Parameter Lightrun’s Approach Distributed Tracing
- Live Instrumentation Allow developers to add Logs, Metrics & Snapshots to anywhere in the code in real-time, on demand Requires upfront span instrumentation and has only limited coverage for auto-instrumentation
- Granularity of Information Get latency measurements and breakpoint-grade information including variables, objects and expression evaluations Get latency measurements and other, partial bits of telemetry
- Developer Experience Instrument directly from the IDE and use actual programming expressions Use dedicated vendor GUIs only and proprietary query languages
- “Step In” vs. “Step Over” Time anything - from a few lines to a full method Time endpoint responses, not application transactions
- Conditional Observability Every action can be filtered by a code-level condition, and defined in real-time No conditionality exists - the user must manually filter pre-defined information
Players: Sentry, Rollbar, Raygun, etc…
Lightrun is a Developer Observability Platform, focusing both on getting to the root cause of issues and on being an active “pane of glass” for any part of the application’s surface, regardless of its specific place in the architecture.
- Comparison Parameter Lightrun’s Approach Exception Monitoring
- Data Focus Any type of code-level data Exceptions and errors
- Timing of Data Collection Information collected before, during and after the exception occurred Information collected during the exception
- Troubleshooting Process Add dynamic logs, metrics and snapshots on the fly, at any time Review existing, static log and context information, only when an exception occurred
- Applicable Issues Suitable for solving a variety of code-level issues, including logical & performance issues Suitable for solving exception-based issues only
- IDE Support Add, consume and analyze telemetry right from the IDE For some platforms, consume limited exception data from the IDE
Players: IntelliJ Remote Debugger, VS Code Remote Debugger, etc…
Debug multi-instance applications, across network boundaries and without ever stopping the live applications. Connect anywhere, anytime without any extra configuration necessary.
- Comparison Parameter Lightrun’s Approach Remote Debugging
- Application Connection No direct application connection required Direct connection via dedicated debug port
- Debugging workflow Add “virtual” breakpoints without stopping execution to get code-level data Stop and start the application execution at every breakpoint to get code-level data
- Performance overhead Capped, throttled performance overhead with a “dead man” switch No performance capping or management
- Side effects Lightrun actions have no side effects by design Every action taken in the session has side effects
- Multi-instance support Add logs, metrics and snapshots to multiple applications at any given time One connection “active” at a time - step in and out of breakpoints in one attached process every time
- Configuration process Immediate connection to the right process, no networking skills required Some networking skills required to find and attach to the right application process
- Security Considerations All access is audited, and Lightrun’s proprietary Sandbox verifies every action on the fly Access is not restricted, and no verification of actions takes place (which can harm the application’s stability and integrity)
Players: JProfiler, jvisualvm, yourkit, Java Flight Recorder, async-profiler, etc…
Lightrun’s Logs, Snapshots and Metrics focus deep, granular data not often available through normal profiling – without incurring the massive overhead associated with profiling.
- Comparison Parameter Lightrun’s Approach Profiling
- Type of SDK Idle, non-intrusive runtime SDK Continuously active, intrusive runtime SDK
- Type of Applicable Issues Deep, granular & specific performance issues as well as logical issues General performance and resource-related issues, no support for solving logical issues
- Volume & Type of Information Produced Brief, concrete application-level telemetry Verbose, long streams of raw profiling information
- Performance Impact Small and sporadic Very heavy and continuous
See Lightrun In Action!
Let us walk you through the platform, one step at a time.