ALB metric equivalent of Latency metric in Classic load balancer

Question

Classic load balancer has 'Latency' metric which we are currently using to scale the instances. We are planning to switch to Application load balancer. What is the metric in ALB that is equivalent of 'Latency' in classic load balancer? is it 'TargetResponseTime'?  
  
How is the performance of ALB against the classic load balancer, saw some of the posts that were complaining about increase in latency after switching to ALB.  
  
Regards,  
Santhosh I C

Answer

For the purposes of scaling EC2 instances, you can consider Classic Load Balancer Latency CloudWatch metric and Application Load Balancer TargetResponseTime CloudWatch metric equivalent. However, you may see that the TargetResponseTimes are much higher as explained below.  
Link:

```
@chris_hansen - There are many different reasons those metrics may show different values. The response time of the application is going to include all requests handled by your application, including internal calls that trigger additional or async transaction while depending on your specific configuration the ELB is only going to record the time spent between the request leaving the ELB and the response being returned. Any additional work done by your application after that response most likely will not be tracked by the ELB metric.
```

ALB metric equivalent of Latency metric in Classic load balancer

Contenuto pertinente