Comment 7 for bug 1978225

Your expectations about the REST API performance are perfectly reasonable, but we need to be transparent on how we are addressing this issue and to create realistic expectations about the time to fix it.

1) why this is not a critical/blocking bug

This is regarded as a product characteristic at this point (a bad one), and we are aware it exposes MAAS to a number of DoS attacks. MAAS API has always been slow, so it's not a regression of the last release, and customers have learned to live with this. Fixing this is not the top priority for the project stakeholders either, so we are not blocking new releases because of this.

2) what's our strategy for improving MAAS performance

During the 3.2 cycle we introduced performance testing and profiling into our CI pipeline. This gave us insight about how the time is spent and what actions would produce the biggest impact. The conclusion was that MAAS needs major overhaul in critical components, which couldn't all be done in a single release cycle. We did what we could in 3.2, and our management is committed to setting aside time on future cycles to make this happen.

3) when MAAS is going to be fast

We'll get there when we get there. There are many unknowns, like how many man-hours we can commit to this each cycle and how fast the MAAS team will grow. We set the goal of getting faster with each release, so you can expect this much.

This might not be the answer you were expecting, but this is the answer we can give you at this time.