After having used Zammad as a classic ticket system for our support staff with good success for over two years now, we recently introduced a separate Dockerized 3.3.0 instance for our inside sales and order processing departments as well. As it turns out, they have very different requirements and I’m not sure how to best handle these for optimal performance.
The situation we’re used to is having a fairly small volume of longer tickets (18k opened in total, 8k remaining after cleanup - deleting spam, merging etc.) and mostly around 3 agents. This means that we were able to evolve a very complex system of triggers and automations in order to make support work easier.
The new ticket system, on the other hand, has seen 1.1k tickets in a single week and has 8 agents in total. These are almost exclusively automated mails required in order processing which need to be linked to each other and to actual customer communication. They will usually be only one article long and be closed quickly.
Unfortunately, I don’t think there’s a clear description of how exactly mail filters, triggers, automations and SLAs are processed and what implications this has for which use cases. But I would assume that we will need to be way more conservative with any of them to start with.
Does someone here have experience with our new use case? More specific questions:
- What to watch out for especially, and how to lighten the load on the system? Keeping overviews few and general was a surprising way to regain a lot of speed in our other instance, so I’m hoping there are more tricks like that.
- Are there any conditions or actions that are especially taxing when using triggers? Is the order of triggers and conditions inside them important, and can we prioritize them in some way?
- Would deleting tickets after e. g. 6 months help or are old tickets not usually relevant for performance?
- What’s the practical maximum of tickets you can work with in a useful way for a given time span?
Thanks for any recommendations!