Application/network level performance testing for MerginMaps CE

[awa5114]

Hi there,

We have recently deployed an instance of MerginMaps CE on a Ubuntu machine using the docker compose file provided in this repo. The instance seems to work OK, but there are some concerns around how it will perform when we scale to around 10 users.

I am wondering whether there are some general guidelines for both network and application level performance testing of MerginMaps CE? I am trying to avoid a situation where we scale but end up hitting performance bottlenecks.

Thanks a lot for your consideration.

[MarcelGeo]

Hi @awa5114 .

We are improving documentation about server. In upcoming release of merginmaps.com/docs are following requirements:

For a typical deployment, we recommend using a dedicated host machine with 8 GB of memory and 2 vCPUS (similar to AWS t3a.large instances).
The requirements for CPU and persistent storage depend largely on the frequency of project updates and the anticipated size of the data you expect to store respectively.
A very conservative rule of thumb, regarding needed disk size would be mergin maps project size * number of versions.
If you have a team size over 25 people and synchronise often your projects, consider a host with 4 vCPUS.

In general, performance is based on size of projects and number of versions.

10 users on standard machine is no problem.

[MarcelGeo]

Any progress here @awa5114 ?