How do I determine optimal sizing for my on premises EUM Server?
This guide provides context to EUM data and recommendations for sizing your on-premises EUM Server for different levels of network traffic (metrics per minute).
After understanding the EUM basic concepts, review the recommended sizing below for your on-premises deployment.
Table of Contents
EUM Basic Concepts
To deploy EUM successfully, you should understand Beacons and Resource Timing Snapshots. Learn more about them in the Understanding EUM and Events Service Concepts article.
EUM Server Sizing Recommendations
The following tables recommend T-shirt sizing (S, M, L) for your on-prem EUM Server based on synthetic load testing. Once you have estimated your EUM traffic profile, you can use these maximum load measurement results to establish which T-shirt size is appropriate for your deployment.
If your EUM traffic profile exceeds the maximum traffic given in the tables, see the Extra Large Sizing Recommendations article.
For simplicity, the sizing recommendations for Browser RUM and Synthetic Monitoring are grouped together. If you are using the EUM Server with the Events Service, see Analytics On-Prem Configuration Recipe Book.
Browser RUM
The sizing suggestions in the table below are based on one web beacon consisting of the following:
- 1 base page
- 7 Ajax requests
- 100 resources
EUM Sizing for Browser RUM
|
EUM
Size
|
Maximum Traffic
(Metrics per Minute)
|
Hardware
Specifications
|
Network
Bandwidth
|
Java
Heap Size
|
Small
|
Beacons |
60k <
|
- CPU: 4 CPU
- RAM: 16 GB
- Hard drive: 300 GB
EC2 Instance Type: m4.xlarge
|
759 Mbps
|
11 GB
|
| Events |
60k |
| Sessions |
3k |
Medium
|
Beacons |
120k |
- CPU: 8 CPU
- RAM: 32 GB
- Hard drive: 300 GB
EC2 Instance Type: m4.2xlarge
|
1000 Mbps
|
28 GB
|
| Events |
120k |
| Sessions |
6k |
Large
|
Beacons |
300k |
- CPU: 16 CPU
- RAM: 64 GB
- Hard drive: 300 GB
EC2 Instance Type: m4.4xlarge
|
2000 Mbps
|
50 GB
|
| Events |
300k |
| Sessions |
12k |
Mobile RUM
The sizing suggestions in the “EUM Sizing for Mobile RUM” table are based on network traffic for iOS and Android applications sending one beacon consisting of the percentage of the data types listed below.
| Data Types |
Percentage of Beacon |
| Network Request |
80% |
| UI Interaction Capture |
10% |
| Screenshots |
1% |
| Touchpoints |
1% |
| Breadcrumbs |
4% |
| ANRs |
1% |
| Code Errors |
1% |
| Crash Reports |
0.02% |
| Custom Timers |
0.01% |
| Custom Metrics |
0.03% |
| System Events |
0.04% |
| Method Calls |
1.9% |
EUM Sizing for Mobile RUM
|
EUM
Size
|
Maximum Traffic
(Metrics per Minute)
|
Hardware
Specifications
|
Network
Bandwidth
|
Java Heap
Size
|
Small
|
Beacons |
90k <
with 5 crashes
|
- CPU: 4 CPU
- RAM: 16 GB
- Hard drive: 300 GB
- EC2 Instance Type:
m4.xlarge
|
750 Mbps
|
11 GB
|
| Events |
70k |
| Sessions |
40k |
Medium
|
Beacons |
90-130k |
- CPU: 8 CPU
- RAM: 32 GB
- Hard drive: 300 GB
- EC2 Instance Type:
m4.2xlarge
|
1000 Mbps
|
30 GB
|
| Events |
70-110k |
| Sessions |
40-61k |
Large
|
Beacons |
130-550k |
- CPU: 16 CPU
- RAM: 64 GB
- Hard drive: 300 GB
- EC2 Instance Type:
m4.4xlarge
|
2000 Mbps
|
50 GB
|
| Events |
110k-3.5M |
| Sessions |
61-118k |
IoT Monitoring
The sizing suggestions below are based on network traffic for IoT applications sending beacons for the following events:
- Network requests
- Custom events
- Error events
EUM Sizing for IoT Monitoring
|
EUM
Size
|
Maximum Traffic
(Metrics per Minute)
|
Hardware
Specifications
|
Network
Bandwidth
|
Java
Heap Size
|
Small
|
Beacons |
110k <
|
- CPU: 8 CPU
- RAM: 32 GB
- Hard drive: 300 GB
|
759 Mbps
|
10 GB
|
| Events |
500k |
Medium
|
Beacons |
110-300k |
- CPU: 8 CPU
- RAM: 32 GB
- Hard drive: 300 GB
|
1000 Mbps
|
22 GB
|
| Events |
500-600k |
Large
|
Beacons |
300-600k |
- CPU: 16 CPU
- RAM: 64 GB
- Hard drive: 300 GB
|
2000 Mbps
|
50 GB
|
| Events |
600k-1M |
Resources
Sizing XL profiles for on premises EUM Server and the Events Service: What's recommended?
Understanding EUM and Events Service Concepts
How do I configure Analytics for on premises Controllers?
