Currently, to raise the alUSD debt cap, the established process is to raise it in million increments (or in terms of around % of current debt cap). This requires us to revisit the approach frequently. Using 2 data points, we have had to raise the cap twice in 10 days (source: me). This is a timely beauraucratic step, and as we see with SLP and curve migrations, devs have to focus on one thing at a time (A.k.a. this is additional dev load).
I see a few options:
- Continue to regroup biweekly, more or less, increasing caps in "bundles" proposed by devs and community
- Remove the alUSD debt cap, releasing the flood gates.
3. Increase the alUSD debt cap as a function of time, instead of manual management.
I will discuss option 3, and initially we can consider increasing the alUSD debt cap by $N million per week. The main question is the proposed shape and slope. Firstly, starting modest seems like the best approach. Secondly, anything is better than nothing. Finally, we do not want to overdo it by allowing too much alUSD mint.
Based on $20 million of alUSD demand going away in 10 days, increasing the cap by $1 million per week linearly, timeboxed in 3 months (12 weeks) would ease customer pain without dev management. We can increase the alUSD mint cap stepwise every hr (my choice), every 24hr, or with each harvest.
Happy to reconsider #s etc., perhaps we're missing something. Any thoughts?
