Customer environment examples

This page contains recommended sizing requirements and related information for typical customer environments.

  • Network - Although proper hardware is one key factor for a smooth working environment, always consider the network of your environment. See network bandwidth best practices for more information.
  • RegEx - Regular Expressions (RegEx) can be very expensive. Certain unintended and intended strings may cause RegEx to exhibit exponential behavior. RegEx is very useful, but developers need to understand how it works and to always be careful with it. For more details, see RegEx - Using Regular Expressions.

Note

In the following examples, the CPU core listings imply the number of physical cores and not virtual cores. Server sizes larger than Medium require a minimum 2.6 GHz.
The suggested memory for the hardware where the server is running is a recommendation. For instance, in a XLarge scenario, 48GB is recommended, because caches and other running processes, and future proof for upcoming releases also need resources.

Small environment

Example

Environment overview

Environment Agents Data centers Transactions Measures Peak Comments
Production Java 5 1 17 million/day 197 tps 2,000 without UEM

Sizing suggestion

Minimal process requirements

Environment Server process Collector process Database process Comments
Production Java 3 GB RAM 4 cores 100 GB local file storage (depends on how long to keep the session store) 2 GB RAM 1 cores 30 GB storage PostgreSQL 1 GB RAM 1 core 100 GB storage

Minimal hardware requirements

Machine/OS Installed components*
x64 Linux or Windows 8 cores 12 GB RAM 1 PostgreSQL database (1 GB heap) 1 AppMon Server (8 GB heap) 1 Small AppMon Collector (2 GB heap)

* All components/processes are running on a single machine.

Minimal storage requirements

Storage Comments
Application and configuration 5 GB
Performance Warehouse storage 5 GB PostgreSQL database is sufficient
Session storage for 1 day 89 GB
Collector application and configuration 30 GB Including class cache

Minimal network requirements

Bandwidth Comments
Network Bandwidth 11 MB Agent <-> Collector

Other considerations:

  • The Memory Analysis Server should be installed to a separate machine if the memory snapshots processed are larger than the remaining physical RAM.
  • A minimum one Collector per physical location (data center); 200 JVMs per collector process.
  • Collectors must be very close to their connected Agents (same LAN).

Medium environment

Example

Environment overview

Environment Agents Data centers Transactions Measures Peak Comments
Production Java 80 1 17 million/day 410 tps 17,000 820 tps without UEM

Sizing suggestion

Minimal process requirements

Environment Server process Collector process Database process Comments
Production Java 6 GB RAM 8 cores 100 GB local file storage (depends on how long to keep the session store) 4 GB RAM 2 cores 50 GB storage PostgreSQL 2 GB RAM 2 core 100 GB storage

Minimal hardware requirements

Machine/OS Installed components*
x64 Linux or Windows 12 cores 16 GB RAM 1 PostgreSQL database (2 GB heap) 1 AppMon Server (6 GB heap) 1 Small AppMon Collector (2 GB heap)

* All components/processes are running on a single machine.

Minimal storage requirements

Storage Comments
Application and configuration 5 GB
Performance Warehouse storage 40 GB PostgreSQL seems basically sufficient. However, Oracle, DB2, or SQLServer database is recommended (especially if you want to extend the environment later). Consult the database provider for database hardware requirements.
Session storage for 1 day 235 GB
Collector application and configuration 50 GB Including class cache

Minimal network requirements

Bandwidth Comments
Network bandwidth 28 MB Agent <-> Collector

Other considerations:

  • The Memory Analysis Server should be on a separate machine if the memory snapshots processed are larger than the remaining physical RAM.
  • A minimum of one Medium Collector per physical location (data center); 200 JVMs per Collector process.
  • Collectors must be very close to their connected Agents (same LAN).

Large environment

Example

Environment overview

Environment Agents Data centers Transactions Measures Peak Comments
Production Java 400 2 52 million/day 1,200 tps 2,400 tps with UEM

Sizing suggestion

Minimal process requirements

Environment Server process Collector process Memory Analysis Server process Collector for monitoring plugins process Databaseprocess Comments
Production Java 12 GB RAM 16 cores 500 GB local file storage (depends on how long to keep the session store) 8 GB RAM 4 cores 50 GB storage 8 GB RAM 2 cores 1 GB RAM 2 cores Oracle or MS SQL 4 GB RAM 4 cores 500 GB storage

Minimal hardware requirements

Machine/OS Installed components
x64 Linux or Windows 16 cores 32 GB RAM 1 AppMon Server (12 GB heap) 1 AppMon Memory Analysis Server (8 GB heap) More heap for analysis server, OS and to be future proof Analysis server is optional. If required, then more cores are needed.
x64 Linux or Windows 6 cores 8 GB RAM 1 Oracle database (4 GB heap)
x64 Linux or Windows 12 cores 24 GB RAM 2 Large AppMon Collectors (8 GB heap each) 1 AppMon Collector for monitoring plugins (1 GB heap)

Minimal storage requirements

Storage Comments
Application and configuration 5 GB
Performance Warehouse storage 191 GB Oracle, DB2 or MS SQL database is needed. Consult the database provider for database hardware requirements.
Session storage for 1 day 349 GB RAID 10 or SAN
Collector application and configuration 50 GB Including class cache

Minimal network requirements

Bandwidth Comments
Network bandwidth 42 MB Agent <-> Collector
Network bndwidth 42 MB Collector <-> Server
Network bandwidth compressed 10.4 MB Collector <-> Server

Other considerations:

  • The Memory Analysis Server should be installed to a separate machine if the memory snapshots processed are larger than the remaining physical RAM.
  • A minimum of one large Collector per physical location (data center); 400 JVMs per collector process
  • Collectors must be very close to their connected Agents (same LAN)

XLarge environment

Example 1

Environment overview

Environment Agents Data Centers Transactions Measures Peak Comments
Production Java 1,000 8 10 million/hour 2,777 tps 200,000 3000 tps 100 nodes per transaction, without UEM

Sizing suggestion

Minimal process requirements

Environment Server process Collector process Memory Analysis Server process Database process Comments
Production Java 22 GB RAM 24 cores >2 TB local file storage (depends on how long to keep the session store) 8 GB RAM 4 cores 50 GB storage 8 GB RAM 2 cores Oracle 8 GB RAM 8 cores 1 TB storage Session storage size depends on how many days to keep the sessions. 3-4 TB should be sufficient for one day. The database load is large, so SSD disks (RAID 1-0) and Performance Warehouse database partitioning are recommended.

Minimal hardware requirements

Machine/OS Installed components
x64 Linux or Windows 24 cores 48 GB RAM 1 AppMon Server (30 GB heap) More heap for OS and to be future proof
x64 Linux or Windows 10 cores 24 GB RAM 1 AppMon Memory Analysis Server (8 GB heap) 1 Oracle database (8 GB heap)
x64 Linux or Windows 16 cores 32 GB RAM 3 Large AppMon Collectors (8 GB heap each)
x64 Linux or Windows 10 cores 24 GB RAM 2 Large AppMon Collectors (8 GB heap each)

Minimal storage requirements

Storage Comments
Application and configuration 5 GB
Performance Warehouse storage 473 GB Oracle database is needed. Consult the database provider for database hardware requirements.
Session storage for 1 day 1,250 GB RAID 10 or SAN
Collector application and configuration 50 GB Including class cache

Minimal network requirements

Bandwidth Comments
Network bandwidth 150 MB Agent <-> Collector
Network bandwidth 150 MB Collector <-> Server
Network bandwidth compressed 40 MB Collector <-> Server

Other considerations:

  • The Memory Analysis Server should be installed on a separate machine if processed memory snapshots are larger than the remaining physical RAM.
  • Collectors must be very close to their connected Agents (same LAN)
  • Installation of fast hard disks such as SSDs
  • Usage of partitioning for database

Example 2

Environment overview

Environment Agents Data centers Transactions Measures Peak Comments
Production Java 1,000 6 800 million/day 9,259 tps (many short UEM paths) 201,000 The 800 million transactions per day for the production environment are the worst case estimate. During the POC, the customer measured the number of transactions for a few high throughput agents, which was then used to compute the expected upper limit.

Sizing suggestion

Minimal process requirements

Environment Server process Collector process Memory Analysis Server process Database process Comments
Production Java 22 GB RAM 24 cores >4 TB local file storage (depends on how long to keep the session store) 8 GB RAM 4 cores 50 GB storage 16 GB RAM 2 cores Oracle 8 GB RAM 8 cores 2 TB storage Do not start with a big bang deployment, but rather start slowly. The expectation is that most of the applications have lower requirements than the measured ones (during the POC). Starting with one server is recommended, perhaps over time adding a second one. The database load is huge, so SSD disks (RAID 1-0) and Performance Warehouse database partitioning is recommended.

Minimal hardware requirements

Machine/OS Installed components
x64 Linux or Windows 24 cores 48 GB RAM 1 AppMon Server (30 GB heap) More heap for OS and to be future proof
x64 Linux or Windows 10 cores 24 GB RAM 1 AppMon Memory Analysis Server (16 GB heap) 1 Oracle database (8 GB heap)
x64 Linux or Windows 16 cores 32 GB RAM 3 Large AppMon Collectors (8 GB heap each)
x64 Linux or Windows 16 cores 32 GB RAM 3 Large AppMon Collectors (8 GB heap each)

Minimal storage requirements

Storage Comments
Application and configuration 5 GB
Performance Warehouse storage 314 GB Oracle database is needed. Consult the database provider for database hardware requirements.
Session storage for 1 day 888 GB RAID 10 or SAN
Collector application and configuration 50 GB Including Class Cache

Minimal network requirements

Bandwidth Comments
Network bandwidth 110 MB Agent <-> Collector
Network bandwidth 110 MB Collector <-> Server
Network bandwidth compressed 27 MB Collector <-> Server

Other considerations:

  • The Memory Analysis Server should be installed on separate machine if the memory snapshots processed are larger than the remaining physical RAM.
  • Collectors must be very close to their connected Agents (same LAN).
  • Installation of fast hard disks such as SSDs.
  • Usage of partitioning for database.

Example 3

Environment overview

Environment Agents Data centers Transactions Measures Peak Comments
Production .NET 500 (w3wp processes) 10 App Pools 1 TB storage 5 80 million/day 926 tps 30,430

Sizing suggestion

Minimal process requirements

Environment Server process Collector processes Database process Comments
Production Java 22 GB RAM 24 cores 8 GB RAM 4 cores 50 GB storage Oracle 8 GB RAM 8 cores 2 TB storage SSD disks (RAID 1-0) and Performance Warehouse database partitioning are recommended. .NET Collectors needs SSD hard disks.

Minimal hardware requirements

Machine/OS Installed components
x64 Linux or Windows 24 cores 48 GB RAM 1 AppMon Server (48 GB heap) More heap for OS and to be future proof
x64 Linux or Windows 8 cores 16 GB RAM 1 Oracle database (8 GB heap)
x64 Linux or Windows 10 cores 24 GB RAM 2 Large AppMon Collectors (8 GB heap each)
x64 Linux or Windows 6 cores 10 GB RAM 1 Large AppMon Collectors (8 GB heap each)

Minimal storage requirements

Storage Comments
Application and configuration 5 GB
Performance Warehouse Storage 48 GB Oracle database is needed. For DB hardware requirements, consult the DB provider.
Session storage for 1 day 89 GB RAID 10 or SAN
Collector application and configuration 50 GB Including class cache

Minimal network requirements

Bandwidth Comments
Network bandwidth 11 MB Agent <-> Collector
Network bandwidth 11 MB Collector <-> Server
Network bandwidth compressed 2.7 MB Collector <-> Server

Other considerations:

  • Collectors must be very close to their connected Agents (same LAN)
  • Max. 250 .NET Agents per Large Collector
  • Installation of fast hard disks such as SSDs
  • Usage of partitioning for database