Good frontend response times do not guarantee your system will perform. Monitor your backend during testing and make sure it does.
Under normal load, frontend resources often account for 80-90% of response time, but as load increases this ratio can change.
Even at load levels most believe* they can handle, backend components become the primary bottleneck and reason for performance degradation.
Used together, performance testing and monitoring give you the meaningful insights you need to optimize your system and locate potentially crippling performance issues before it's too late.
Two ways to monitor your servers with Load Impact
Option 1: Insert your New Relic API key in our platform and begin analyzing server and test data in one place, immediately. And since we don't roll-up data, you can keep that analysis forever.
Option 2: Download and install our Python based monitoring agents on your machines - packaged for Linux and Windows. Open source code available on GitHub.
- CPU usage
- Memory usage
- Disk I/O, bytes/s, read bytes/s, write bytes/s, usage percentage
- Network I/O, bytes/s, received bytes/s, sent bytes/s, packets/s, received packets/s, sent packets/s Basically, if it can be extracted/calculated/printed/pulled from somewhere in any way — it can be collected as a metric.
Metrics provided by installed Load Impact monitoring agents:
- CPU usage
- APDEX This is just the beginning, we'll be adding more New Relic metrics regularly. Let us know what you need.
Metrics provided by New Relic APM integration:
Ready to start monitoring your servers?