A few months back, our team moved DevCentral.f5.com from our corporate datacenter to a cloud service provider. As part of this project, we were required to take over the tasks previously performed by our IT department. One of the tasks that fell into my hands was to look at the monitoring and alerting for the heath of the systems in our environment. These systems included our application tier (application, database, and storage) as well as the F5 network infrastructure (GTM, LTM, ASM, WA, APM, etc).
After looking at several product offerings, we ended up choosing PRTG Network Monitor from Paessler. We chose PRTG for it’s cost model as well as it’s ability to be customized and extended.
This is the first in a series of articles covering PRTG and how we use it to monitor the production environment of DevCentral.f5.com. In this article, I’ll go over the features in the product, the installation process, and touch on configuration. In future articles, I’ll talk in more detail about configuration and the development of custom monitoring (to get data from our F5 products) and alerting (sending out customized email/SMS notifications of the monitor results).
PRTG
The PRTG Network Monitor runs as a service on a Windows machine on your network. It collects statistics from the components in your network (which it can auto-discover) and retains that data for historical performance reporting and analysis. The product has the following features:
- Quick Download, Installation, and Configuration
- Choose Between 5 Easy To Use User Interfaces
- Comprehensive Network Monitoring with more than 130 sensor types covering all aspects of network monitoring.
- Flexible Alerting
- Support for Clustered configuration of Network Monitors.
- Distributed Monitoring Using Remote Probes
- Data Publishing and Maps
- In-Depth Reporting
Installation
You can download the install package from the product website. It consists of a self-contained installer that does it all for you. Since PRTG uses it’s own internal database, no 3rd party products are required to use the software.
I ran the install with all the defaults and within a few minutes, the software was ready to use. When the install finished, a web browser was opened with the Configuration page. After configuring a few things I was ready to start adding sensors to my network gear. I opted to not do the auto-scanning of my network but went into the Sensors page by clicking on the “Review Results” button.
Sensor Configuration
- Root
- Local Probe (I’m only using single probe so everything under this item is monitored locally).
- management server (this is the local machine PRTG is running on. PTRG can monitor itself by the way).
- DC1
- F5 Devices
- asm-1 (device 1 in our Application Security/Local Traffic HA pair)
- asm-2 (device 2)
- egw-1 (device 1 in our Edge Gateway/Web Acceleration HA pair)
- egw-2 (device 2)
- gtm-1 (our Global availabililty DNS device)
- Virtuals
- devcentral.f5.com (external virtual for application monitoring)
- Servers - DOMAIN
- adc-1 (windows domain controller #1)
- adc-2 (#2)
- Servers - APPS
- app-1 (Application server #1)
- app-2 (Application server #2)
- … (Application server #n)
- db-1 (Database server #1)
- db-2 (Database server #2)
- fil-1 (Network file storage)
- F5 Devices
- Local Probe (I’m only using single probe so everything under this item is monitored locally).
Each of those tree items allows for multiple sensors attached to them. I made use of memory, CPU, disk, Database, and system health checks.
In the next article in this series, I’ll break down the various health checks I used and cover a few custom ones we developed to determine our application’s health.