Performance / volume estimates


#1

Hi all
Are there any estimates of what volume can RT (in my case RTIR) handle? We are sending lots of alerts from time to time to our RTIR and I would like to know if anybody has run any stress tests related to the volume.


#2

Running our RT installation for 14 years now, currently 102.000 tickets, increasing by 50-500 tickets per day. Virtual machine with 2 cores and 6GB of RAM is more than enough, even full text search returns results after few seconds.

Bye, Daniel


#3

Thank you Daniel
It’s not the number I’m looking for. We’re currently running our RTIR instance with ~15.000 automatically generated incident reports per week, and another 3.000 - 4.000 additional tickets (incidents and countermeasures). That’s why I asked for stress tests. As you realize we reached your current ticket volume in just 2 months.


#4

Well I’ve seen RT instance with 1,5M of tickets, thousands of users, hundreds of queues, hundreds of Custom fields and hundreds of groups. It was performing quite well but needed some tuning (mostly DB but also some queries). The complexity grows as cube, not linear - if you add a queue, the performance is influenced also with other already defined things (groups, CFs, …). If you’ve got only the huge amount number of tickets, it should be fine. Only one dimension of the whole cube is being stressed.

Next it depends on the database engine, I’d prefer PostgreSQL for big instances rather than MySQL/MariaDB. MariaDB is better than MySQL now so if you need to stay in that world, you should consider moving to Maria. Also the hardware counts but AFAIK the software itself scales quite well.
Petr


#5

Thank you, that is good information. I still don’t know what are the spikes it can manage, but it’s still good information.