New Relic APM Checklist: 10 Things Before Going to Production
I’ve seen 3 production agent deployments fail this month. All 3 made the same 5 mistakes. It’s clear we need a solid New Relic checklist before heading into production.
1. Set Up Transaction Tracing
This feature allows you to analyze specific requests and transactions within your applications. It matters because if you can’t see what’s happening in detail during a request, you’ll struggle to identify performance bottlenecks.
newrelic.config={ "transaction_tracer.enabled": true }If you skip this, you’ll miss crucial performance insights, which could lead to slow response times and unhappy users.
2. Configure Custom Metrics
Custom metrics help you track application-specific events that New Relic doesn’t capture out of the box. Understanding how your app performs in unique scenarios matters. Otherwise, you’re left in the dark about critical application events.
import newrelic.agent
newrelic.agent.record_custom_metric('Custom/Metric/Name', value)Neglecting this means you’re not capturing data that could be critical to your business needs. You could make decisions based on incomplete data.
3. Define Alert Policies
Setting up alert policies is mandatory. You want to know when things go sideways. New Relic has decent default alerts, but customizing them to your specific app can make a difference.
newrelic/alarm_policy={ "conditions": [ {"name": "High Response Time", "threshold": 300} ] }Skip this, and you risk being blindsided by performance drops that could affect revenue and user satisfaction.
4. Instrument Key Transactions
Key transactions are the lifeblood of your application. Identifying and instrumenting them lets you focus your performance tuning efforts. If not done, you’re essentially flying blind.
newrelic.agent.set_transaction_name('Your/Key/Transaction')Failure to do this means you won’t know how the most critical paths in your app are performing, potentially leading to major issues.
5. Enable APM Features for Web Apps
If you’re building a web application, turning on APM features like real user monitoring (RUM) should be at the top of your priority list. It offers insights into the end-user experience, which is crucial for identifying frontend bottlenecks.
newrelic.config= { "browser_monitoring.auto_instrument": true }If you skip it, you’ll lack visibility into how users actually experience your app, which could ruin your reputation.
6. Integrate with Incident Management Tools
Integrating New Relic with incident management systems like PagerDuty helps streamline your response to alerts and incidents. This matters because if your team doesn’t get alerts in real-time, issues can fester longer than they should.
newrelic/integration={ "service": "PagerDuty", "api_key": "YOUR_API_KEY" }Without this integration, you’ll face longer downtime, impacting both user experience and your support team’s productivity.
7. Set Up Logging Integration
Integrating logs with New Relic enhances your ability to troubleshoot issues in real-time. If logs are separate, solving problems becomes a logistical nightmare.
from newrelic import agent
agent.log('message', 'debug')Skipping this can lead to inefficient debugging and longer resolution times during outages.
8. Limit Data Retention Period
This is often overlooked. You don’t need to keep every bit of data forever, as it can clutter your insights and increase costs. Setting a reasonable data retention policy saves you money and streamlines your data access.
newrelic.config={ "data_retention": 90 }If you forego this, you risk incurring unnecessary costs and overwhelming your team with too much data.
9. Document Your Configuration
For anyone involved in managing your New Relic setup, a clear documentation of configurations is key. It allows for quick troubleshooting and smooth onboarding of new team members. This may seem trivial, but it saves time and headaches down the line.
If overlooked, your team may face confusion and misconfigurations that lead to suboptimal monitoring performance.
10. Regularly Review Performance Metrics
It’s easy to set things up and then forget about them. Regular reviews of your performance metrics help identify areas of concern before they become chronic issues.
newrelic.metrics.list()By not reviewing performance regularly, you’ll miss shifts in resource utilization or other trends that could lead to systemic issues.
Prioritization of the Checklist
- Do This Today:
- Set up transaction tracing
- Define alert policies
- Instrument key transactions
- Enable APM features for web apps
- Nice to Have:
- Configure custom metrics
- Integrate with incident management tools
- Set up logging integration
- Limit data retention period
- Document your configuration
- Regularly review performance metrics
Tools and Services
| Task | Tool/Service | Free Option |
|---|---|---|
| Transaction Tracing | New Relic APM | Yes, limited |
| Custom Metrics | New Relic APM | Yes, limited |
| Alert Policies | New Relic APM | Yes, limited |
| Incident Management | PagerDuty | Yes, limited |
| Logging Integration | Loggly | Yes, basic |
| Data Retention Management | New Relic APM | Yes, limited |
The One Thing
If you only do one thing from this list, make sure to set up transaction tracing. Here’s why: understanding detailed transaction flow is essential for diagnosing performance problems. Without this feature, you’re making decisions blindfolded – risky and stupid.
FAQ
- What’s the best way to learn New Relic?
There’s no replacement for hands-on experience. Set up a free account and play around with it. - Can I use New Relic with any programming language?
Yeah, New Relic supports a wide array of languages, from Ruby to Python to Java. - How often should I review my New Relic settings?
Quarterly is a good baseline, but if you have a dynamic application, consider monthly reviews. - Where can I find more resources?
Check out the New Relic Docs and the New Relic Community Forum.
Data Sources
- New Relic Documentation
- Community forums and benchmarks
Last updated April 21, 2026. Data sourced from official docs and community benchmarks.
🕒 Published: