Is Cacti the right solution for us?

[Nessnaj]

We are thinking about starting to use Cacti. we are researching if Cacti is the right product. Can you give us answers if Cacti meets our requirements?

F2 - Incident reports must be available by e-mail and / or app

F1 - System must be able to run on a virtual server environment

F12 - Both Windows and Linux systems and all equipment with a network component must be able to be monitored

F9 - Alarms must have clear gradations (green, yellow, orange and red)

NF2 - The network monitoring system requires no more than 1 hour of maintenance work per week
F13 - Servers should be easily linked to the network monitoring system using binaries and / or a package manager

F14 - Reporting possibility for viewing up-and downtime of servers

F10 - Standard templates / templates for applying configurations should be available

F18 - The system must be at least available in the English language

NF3 - One-time investment in software and no monthly fees or bound to contract

F15 - Reporting possibility overview of incidents per time span

NF1 - Network monitoring system must be open source

NF4 - Implementation time cannot exceed 24 working hours

F4 - Clear dashboard indicating status of fully monitored ICT environment

F16 - Possibility of a live status overview that clearly shows when incidents occur.

F8 - Historical data should be cleaned up automatically by the system after a set historical date

Thnx in advance

[netniV]

The following is my experience as both a user and a new developer to Cacti. It's not exhaustive and some of my colleagues with more experience will likely have a better view on this.

F2 - Incident reports must be available by e-mail and / or app

Using the THold plugin, this is possible.

F1 - System must be able to run on a virtual server environment

It can run in a virtual environment, but it can be processor/disk intensive depending on how many hosts you are monitoring. I'm monitoring 134 hosts with 1842 graphs on a small VMware box quite happily with 1 minute polling. But that can also depend on what templates you are using to produce your graphs and now many graphs you have.

However, I do know that we have people monitoring thousands of hosts using Cacti. When it comes to a larger scale, most monitoring systems will expect you to use a physical machine rather than a virtual one to reduce the overheads between processor, disk and software.

F12 - Both Windows and Linux systems and all equipment with a network component must be able to be monitored

As with all monitoring systems, that is open ended. Cacti utilises both SNMP and custom scripts to be able to monitor anything, it just depends on what you want to monitor, if someone has done that already and if not, how much effort you are willing to put into customising things for yourself.

F9 - Alarms must have clear gradations (green, yellow, orange and red)

Thold does have a coloured scale for it's statuses.

NF2 - The network monitoring system requires no more than 1 hour of maintenance work per week

I don't do anything more than update the source every once in a while. It pretty much takes care of itself.

F13 - Servers should be easily linked to the network monitoring system using binaries and / or a package manager

Cacti has both network automation scanning, and the concept of remote pollers (which provide downstream polling of devices for more effeciency).

F14 - Reporting possibility for viewing up-and downtime of servers

There is the built in Reporting facility to generate email reports and an additional plugin called ReportIT that can generate aggregate reports.

F10 - Standard templates / templates for applying configurations should be available

Cacti has the concept of Device, Data and Graph templates.

F18 - The system must be at least available in the English language

Cacti is multi-lingual though not all translations are available as any foreign language translations are community provided.

NF3 - One-time investment in software and no monthly fees or bound to contract

Cacti is Open Source Software

F15 - Reporting possibility overview of incidents per time span

Whilst you can see what incidents occurred, I'm not sure if you can generate a report of them.

NF1 - Network monitoring system must be open source

Answered above.

NF4 - Implementation time cannot exceed 24 working hours

This would all depend on what you decide to do. You can have Cacti up and running in less than hour, but as you use the system (as with any new system) knowledge will grow and change of what it is capable of.

F4 - Clear dashboard indicating status of fully monitored ICT environment

This would all depend on what you want from the Dashboard. There are two plugins which provide different "dashboard" style monitoring. Intropage checks and verifies various elements of Cacti, such as problematic hosts, bad logins, top 5 worst availability, Thresholds status (from thold) and events. Monitor provides a view of all hosts with thresholds and the status of them.

F16 - Possibility of a live status overview that clearly shows when incidents occur.

see above

F8 - Historical data should be cleaned up automatically by the system after a set historical date

This happens depending on the profile being used for the data source.

____

Please remember that these are my personal views on the software as a user/admin/developer. Is Cacti right for you? I think it would work, it's free and the community will do what they can to help you via these forums and if there are actual bugs, we will try to fix this quickly via our GitHub issues.

[Nessnaj]

Thanks for your answers!

[amitayb]

in addition to netniV...
I'm using version 1.1.38 willing to upgrade to 1.2

I'm monitoring 2400 devices with total 153,000 items - most of them SNMP and some py scripts (~2500).
I use 2 servers for this mission.

my total poll time is between 45 - 62 seconds!

feel free to ask

Amitay.