Best practices, optimization & learning resources

APM Insight delivers the most value when it's used continuously—not just during incidents. This document helps you move from reactive troubleshooting to proactive performance optimization, using APM Insight's features the way they were designed to be used.

Table of contents

• Best practices for effective monitoring in APM Insight
• Using Apdex scores effectively
• Systematic performance optimization process
• Use Milestones to measure optimization impact
• Reading transaction traces effectively
• When to use custom instrumentation
• Alert configuration strategy
• Enterprise deployment best practices
• Learning & support resources

1. Best practices for effective monitoring in APM Insight

Focus on business-critical transactions

Not all transactions deserve equal attention. In APM Insight, prioritize the transactions that directly impact revenue, customer experience, or SLAs.

How to do this in APM Insight

1. In the APM Insight monitor, click on the Transactions tab
2. Identify transactions with:
   • High throughput
   • High response time
   • Direct business impact (checkout, login, payment)
3. Mark them as Key Transactions
4. Configure alerts and dashboards specifically for these transactions

Why this matters

Optimizing a rarely used admin API won't move the needle—but improving checkout latency by 500 ms can directly impact conversions.

Expected outcome

• Faster troubleshooting
• Alerts that matter
• Clear focus for optimization efforts

Monitor trends, not just spikes

A single spike might be noise. Gradual degradation is often more dangerous.

What to track in APM Insight

• Response time trends (7-day and 30-day averages)
• Apdex score trends
• Database query count growth
• Increase in external service latency

Where to look

• Overview page → Trend graphs
• Reports → Historical performance reports

Why this matters

A response time increase from 1.2s to 1.8s over two weeks may not trigger alerts—but it signals growing user frustration.

Expected outcome

Catch performance degradation weeks before users complain, enabling planned optimization instead of firefighting.

Fine-tune monitoring behavior using Agent Configuration Profiles

Why this matters

Different applications have different monitoring needs. The APM Insight Agent Configuration Profile allows you to fine-tune how web and background transactions are monitored by adjusting key configuration parameters.

This helps balance monitoring depth with operational stability—especially in high-throughput or production environments.

Best practice

Use Agent Configuration Profiles to control monitoring behavior centrally at the application level, ensuring consistent configuration across all application instances.

What you can configure

• Threshold values for web transactions
• Threshold values for background transactions
• Agent-specific monitoring parameters based on platform (Java, .NET, Node.js, etc.)

When configuration tuning is useful

• Applications with high transaction volume
• Environments where default thresholds are too aggressive or too relaxed
• During optimization efforts, to focus on specific transaction types
• To standardize monitoring behavior across multiple instances

Where to configure

1. Navigate to Settings → Discovery & Data Collection → APM Insight Agent Configuration Profile
2. Create, clone, or edit profiles
3. Associate the required profile with your application

Expected outcome

• Consistent monitoring behavior across all instances
• Better control over transaction tracking thresholds
• Reduced noise from non-critical transactions
• Monitoring that aligns with real application usage patterns

Integrate RUM for complete application visibility

Why this matters

APM Insight monitors server-side performance—application code, databases, and APIs. However, users experience the entire journey, from the server to the browser.

Client-side factors such as JavaScript execution, browser rendering, network latency, and asset loading can add several seconds to page load time—even when the server is fast.

The blind spot

Your server may respond in 500 ms, but if the page takes 4 seconds to fully load in the browser, users will still perceive the application as slow.

APM Insight alone cannot explain these client-side delays.

Best practice

Integrate Real User Monitoring (RUM) with APM Insight to achieve end-to-end visibility, from backend processing to actual user experience.

When RUM Integration is critical

RUM is especially important if your application is:

• Customer-facing (e-commerce, SaaS platforms)
• JavaScript-heavy (React, Angular, Vue)
• Used by a global audience with varied network conditions
• Mobile-responsive or mobile-first
• Showing good server metrics but still receiving user complaints

Quick Setup (≈5 minutes)

1. Navigate to your APM Insight monitor
2. Open the RUM Analytics tab
3. Select the appropriate RUM monitor from the dropdown
4. Click Associate

What You Gain

By combining APM Insight with RUM, you get:

• Client-side page load and rendering metrics
• JavaScript errors impacting user experience
• Performance insights by browser, device, and geographic location
• Unified server-side and client-side Apdex scores

How to interpret APM + RUM together

Daily Monitoring Routine

1. Review the APM Insight dashboard for server health
2. Check RUM Analytics for client-side experience
3. Compare server-side and client-side Apdex scores to identify gaps

Expected Outcomes

Short-term (first month):

• Visibility into previously hidden client-side issues
• Faster identification of server vs. client problems
• Better prioritization of optimization efforts

Long-term (6+ months):

• Improved user experience through targeted frontend optimizations
• Fewer "application is slow" support tickets
• Proactive detection of performance regressions
• Higher conversion rates and overall user satisfaction

✅ Do:

• Monitor both server-side and client-side Apdex scores
• Check RUM when server metrics look healthy but users complain
• Set alerts for both backend and frontend performance

❌ Don't:

• Assume fast server response equals good user experience
• Ignore RUM data just because APM metrics look fine
• Use RUM only during incidents—monitor proactively

For detailed RUM integration steps, refer to Document 1: Getting Started with APM Insight.

2. Using Apdex scores effectively

Setting Apdex thresholds that match business reality

Recommended starting thresholds:

How to set Apdex in APM Insight

1. Start with the industry thresholds above
2. Monitor performance for 1 week
3. Adjust thresholds:
   • If Apdex < 0.7 → loosen thresholds temporarily
   • If Apdex > 0.9 → tighten thresholds to drive improvement
4. Review quarterly

Why this matters

Apdex directly reflects user satisfaction, not just technical performance.

Expected outcome

• Clear, measurable UX goals
• Fewer subjective performance debates

3. Systematic performance optimization process

Follow this monthly optimization workflow using APM Insight.

Week 1: Identify top opportunities

Database tab

Look for queries with:

• Total execution time > 5% of total DB time
• Execution time > 100 ms
• Execution count > 10,000/hour

Transactions tab

Focus on transactions with:

• Response time > 2× Apdex threshold
• Throughput > 1,000 requests/hour

Service map

Identify services with:

• Response time > 500 ms
• Increasing error rates

Week 2: Analyze & prioritize

• Use Transaction traces to find causes
• Prioritize based on:
  • Apdex impact
  • Business criticality
  • Effort vs reward (quick wins first)

Week 3: Implement & validate

Optimize:

• Add DB indexes
• Refactor inefficient methods
• Introduce caching

Monitor for 48—72 hours

Compare before/after metrics:

• Response time
• Apdex
• Throughput

Week 4: Document & prevent regression

• Document changes
• Share results with stakeholders
• Add alerts to prevent recurrence

4. Use Milestones to Measure Optimization Impact

Best practice: Create milestones before and after optimization efforts to clearly quantify improvements and detect performance regressions early.

Recommended Milestone Cadence

For every deployment

• Create a pre-deployment milestone to capture the baseline
• Create a post-deployment milestone (after 24—48 hours)
• Compare metrics to quickly identify regressions or gains

Monthly performance reviews

• Create a milestone on the first day of each month
• Track month-over-month trends
• Spot gradual performance degradation or steady improvement

Before major optimizations

• Database query optimization → capture baseline first
• Code refactoring → document current performance
• Infrastructure upgrades → establish a comparison point

After major traffic events

• Product launches, campaigns, or peak traffic periods
• Compare performance under load vs. normal conditions
• Use insights to guide capacity and scaling decisions

Example: Quarterly Optimization Using Milestones

Result: Clear, data-backed evidence of optimization impact

Milestone Best Practices

✅ Do:

• Use descriptive names (e.g., "Pre v2.4 Deployment")
• Always create milestones before making changes
• Wait 24—48 hours after changes before comparing
• Document what changed between milestones
• Share milestone comparisons with stakeholders to demonstrate ROI

❌ Don't:

• Create too many milestones (limit: 50 per application)
• Compare milestones during abnormal traffic patterns
• Ignore gradual negative trends across milestones
• Create milestones without clear context or purpose

Interpreting Deviation Percentages

Milestone deviation highlights how metrics changed between time periods.

Positive deviations (improvement):

• Apdex increase: +15% (0.75 → 0.86) ✅
• Response time decrease: —30% (2.0s → 1.4s) ✅
• Error rate decrease: —50% (4% → 2%) ✅

Negative deviations (regression):

• Apdex decrease: —12% (0.88 → 0.77) ❌
• Response time increase: +45% (1.5s → 2.2s) ❌
• Error rate increase: +80% (1% → 1.8%) ❌

How to respond to regressions

1. Identify what changed between milestones (deployment, config, traffic)
2. Review APM Insight transaction traces for the affected period
3. Roll back or fix if the regression is severe
4. Create a new milestone after remediation to confirm recovery

5. Reading transaction traces effectively

How to identify disproportionate execution time

In the Transaction Trace view, look for:

• Methods consuming >30% of total transaction time
• Repeated method calls
• Queries executed multiple times per request

6. When to use custom instrumentation

Use custom instrumentation for:

• Legacy modules
• Business-specific workflows
• Third-party libraries not auto-instrumented
• Gaps in trace visibility

Where to configure

Use the Custom Instrumentation configuration in APM Insight.

Expected outcome

Complete visibility into hidden performance bottlenecks.

7. Alert configuration strategy

Two-tier alert system

Tier 1 — Warning alerts

(Investigate during business hours)

• Response time > 3s for one or more consecutive 5-minute cycles
• Apdex score < 0.75 for two consecutive 5-minute cycles
• Error rate > 2% for one or more consecutive 5-minute cycles

Tier 2 — Critical alerts

(Immediate action required)

• Response time > 5s for two consecutive 5-minute cycles
• Apdex < 0.5 for one or more consecutive 5-minute cycles
• Error rate > 10% for one or more consecutive 5-minute cycles

Alert fatigue prevention

• Rely on sustained 5-minute evaluation windows instead of instant spikes
• Configure maintenance windows during deployments and planned changes
• Review alert thresholds monthly and adjust based on real-world incidents
• Disable or tune alerts that do not lead to clear actions

8. Enterprise deployment best practices

For multi-team environments

• Assign monitors by application ownership
• Configure team-specific alerts
• Create role-based dashboards

For distributed & microservices systems

• Use Service Map to track dependencies
• Monitor each service as a separate instance
• Enable cross-application visibility

For high-traffic applications

• Enable sampling where applicable
• Use different data retention for:
  • Business-critical apps
  • Internal tools
• Leverage reports for capacity planning

For compliance & audits

• Document monitoring coverage
• Use reports to demonstrate SLA adherence
• Configure retention policies as required

9. Learning & support resources

• APM Insight overview documentation
• Agent configuration guides
• Knowledge base & troubleshooting articles
• Video walkthroughs & demos

Final outcome

By following these best practices, teams using APM Insight can:

• Prevent performance issues instead of reacting to them
• Align monitoring with business impact
• Scale observability across teams and environments
• Continuously improve application performance with confidence