Hardware requirements for Zammad 7

Technical assistance
1

Infos:

  • Used Zammad version: 7.*
  • Used Zammad installation type: package or docker-compose
  • Operating system: Ubuntu 24 LTS
  • Browser + version: different

Hello everyone,

I would appreciate it if you could answer the following three questions as soon as possible. Thank you!

Question 1:

We plan to implement Zammad in our production environment within the next two months. Based on various preliminary tests, the following questions have arisen. We’re assuming 15 agents. On average, we handle approximately 5,000 tickets per year (and this number is increasing). How much hard drive space, RAM, and CPU should we allocate to ensure we’re on the safe side for the next 10 to 15 years? I understand that there’s no single answer to this, as it makes a difference whether a ticket contains only text or many attachments. However, perhaps someone could take a look at their system and give us a rough estimate. We want the system to not only be functional but also to run smoothly! Since it will be running virtually, we should consider whether to place the PostgreSQL database on its own dedicated disk system. I’d like to create a second test instance from a snapshot later on, once there’s a lot of data, so I can test it before deploying it back to production. This also requires factoring in buffer time. How do you handle such test scenarios?

Question 2:
If you’re self-hosting and going into production, what are the pros and cons of installing Docker?

Question 3:

Are there any Zammad Rail commands that can and should be run daily via cron job for maintenance? If so, which ones and why?

Kind regards
H-BLOGX

2

Hi,

Absolutely not, no. Everything that needs to run on a regular basis, is run by Zammad.

Sorry, but that’s very individual (e.g. your skill level in the team). Tech wise there shouldn’t ™ be any big differences.

Nobody will be able to tell you about your storage needs, ever. That’s why even the documentation clearly gives no recommendation. Adding further storage usually also isn’t that big of a problem on non dedicated machines. You wanna pin ball it? Sure, estimate the average email / ticket size you receive, average that over the timespan and add a error margin of your choice, e.g. 50% on top. Then you have a rough estimate. Again, no exact numbers possible. It highly depends on the instance, data and how repeditive the data is.

For the rest of the hardware, please take a look at the documentation:

It has a recommendation for up to 40 agents. Once more, it always depends on how you work etc etc. Should be good enough.