# Core Infrastructure Code Stress Tests

**URL:** <https://forum.mergetb.net/t/core-infrastructure-code-stress-tests/39>\
**Category:** RFCs\
**Created:** [April 22, 2020, 3:19pm UTC](https://forum.mergetb.net/t/core-infrastructure-code-stress-tests/39 "2020-04-22T15:19:01Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![ry](https://forum.mergetb.net/user_avatar/forum.mergetb.net/ry/32/64_2.png) [@ry](https://forum.mergetb.net/u/ry)\
**Post date:** [April 22, 2020, 3:19pm UTC](https://forum.mergetb.net/t/core-infrastructure-code-stress-tests/39/1 "2020-04-22T15:19:01Z")

</div>

# Core Infrastructure Code Stress Tests

We need stress tests that can pound on our core infrastructure code without having to construct a full fidelity VTE to do so. For example, in DCompTB we need to be able to scale up to 1k node experiments. Clearly we cannot do this in a full fidelity VTE environment. However, we need to understand where the choke points will be before actually running into scale walls.

etcd is a good example of something we can test at scale for large experiments without having to actually have 1k VMs in a a VTE. We know we are going to hit walls with etcd. So if we can somehow mock up the VM touch points with the cogs then we can test etcd/MinIO/etc at scale.

## Proposal

Create a VTE-like environment where the nodes are mocked up. It would be nice to keep switches but not totally necessary.