Sentry Insights

Sentry Insights provides comprehensive performance monitoring for your applications, helping you identify and resolve performance bottlenecks in both API endpoints and page loads.

Overview

SkyU's Sentry integration offers real-time performance metrics powered by Sentry's monitoring capabilities. The Insights dashboard gives you visibility into:

API endpoint response times
Page load performance
Performance trends over time
Environment-specific metrics

Before using Sentry Insights, ensure you have configured your Sentry integration in your project settings.

Performance Metrics

API Calls

Monitor your backend API performance with detailed metrics:

Metric	Description	Use Case
P50 (ms)	Median response time	Understand typical user experience
P95 (ms)	95th percentile response time	Identify worst-case performance
Endpoint URL	API endpoint path	Track specific endpoints

Page Loads

Track frontend performance for your web applications:

Metric	Description	Use Case
P50 (ms)	Median page load time	Measure typical page performance
P95 (ms)	95th percentile load time	Catch slow page loads
Page Route	Frontend route path	Monitor specific pages

Using Filters

Sentry Projects

Select one or multiple Sentry projects to focus your analysis:

View performance across all projects
Compare metrics between different applications
Isolate issues to specific projects

Environments

Filter by environment to analyze performance in different contexts:

Production: Monitor real user performance
Staging: Validate changes before deployment
Development: Track performance during development

Stats Period

Choose the time range for your analysis:

24h: Recent performance trends
7d: Weekly patterns
14d: Two-week overview
30d: Monthly trends
All Time: Historical data

P95 Duration Threshold

Filter endpoints or pages by minimum P95 response time:

Focus on slow endpoints exceeding your performance budget
Prioritize optimization efforts
Track improvements after optimization

Interpreting Results

Understanding Percentiles

⚠️

Why P95 matters: While P50 shows the median experience, P95 reveals performance issues affecting 5% of your users. For a high-traffic application, this could represent thousands of users experiencing slow performance.

Example:

P50 = 200ms: Half of your requests complete in under 200ms
P95 = 1500ms: 5% of requests take longer than 1.5 seconds

Performance Benchmarks

General guidelines for web performance:

Response Time	User Experience	Action
< 100ms	Instant	Excellent
100-300ms	Slight delay	Good
300-1000ms	Noticeable	Acceptable
1000-3000ms	Slow	Needs improvement
> 3000ms	Very slow	Critical issue

Common Use Cases

1. Identifying Performance Bottlenecks

Filter by High P95

Set the P95 duration filter to identify slow endpoints (e.g., > 1000ms)

Analyze Patterns

Look for common characteristics:

Specific endpoints or pages
Particular environments
Time-based patterns

Prioritize Fixes

Focus on endpoints with:

Highest P95 values
Most frequent usage
Critical user paths

2. Validating Optimizations

Baseline Measurement

Record P50 and P95 values before optimization

Deploy Changes

Release your performance improvements to staging or production

Compare Metrics

Use the same time period and filters to measure improvement

Monitor Trends

Track metrics over several days to ensure consistent improvement

3. Environment Comparison

Compare performance across environments to:

Validate that staging matches production performance
Identify environment-specific issues
Ensure optimizations work in all environments

Best Practices

Regular Monitoring

Check Sentry Insights daily for production environments
Set up alerts for performance degradation (via Sentry)
Review weekly trends to catch gradual performance decline

Performance Budgets

Establish performance budgets for your application:

API Endpoints:
- P50 < 200ms
- P95 < 500ms

Page Loads:
- P50 < 1000ms
- P95 < 2000ms

Use the P95 duration filter to quickly identify violations.

Investigation Workflow

When you identify a slow endpoint:

Check Sentry: Click through to Sentry for detailed traces
Review Code: Examine the endpoint implementation
Check Dependencies: Look for slow database queries or external API calls
Test Locally: Reproduce the issue in your development environment
Optimize: Implement improvements (caching, query optimization, etc.)
Validate: Use Sentry Insights to confirm improvement

Cross-Team Collaboration

Share insights with your team:

Developers: Identify code-level optimizations
DevOps: Investigate infrastructure issues
Product: Prioritize performance improvements
QA: Validate performance in testing

Troubleshooting

No Data Showing

If you don't see any data in Sentry Insights:

Verify Sentry integration is configured correctly
Ensure your application is sending performance data to Sentry
Check that you've selected the correct Sentry project
Verify the selected time period has data

Unexpected Performance Metrics

If metrics seem incorrect:

Verify you're looking at the correct environment
Check for recent deployments that might affect performance
Review Sentry directly for detailed transaction traces
Ensure your application's Sentry SDK is up to date

Missing Endpoints or Pages

If expected endpoints don't appear:

Ensure Sentry is configured to track performance for those endpoints
Verify the endpoint has received traffic in the selected time period
Check that the P95 duration filter isn't excluding the endpoint
Confirm the endpoint is part of the selected Sentry project