We’re going to continue as is, but may not get down to 3 regions. Plan is to go slowly - spin one down, monitor latency and cost, and make a decision to continue.
Alex did believe there were some regions that could be spun down, but not as many as we believed.
Our Team has been focused heavily on the Portal v2 work and QoS over any regional reductions. We have found cost savings in other areas for the time being that are satisfactory and have put this initiative on hold.
I can provide more information as we get back to this workload around methodology, key metrics, etc.
You are not going to retire any portal then?
Not even those that are really close? Like keep us-west-1 close us-west-2 or keep eu-west-1 close eu-west-2/eu-west-3 or keep us-east-1 close us-east-2.
QoS? are you planning to change the Cherry Picker process? is there any place I can follow this?
We will re-evaluate our approach to regions and any reductions after we are at a comfortable position with the development of Portal V2.
No changes to the V1 Cherrypicker will be made. When we release the Portal V2 we will be sure to clearly document and explain the QoS Module that has been implemented.
They are both elements of the Pocket Portal which is separate and distinct from the actions of the Pocket Protocol (v0). The Pocket Portal is essentially an application layer on top of the Pocket Protocol.
The “Cherrypicker” (which is a polarizing name) is part of the current Portal (v1) logic. We are currently in development on the Portal V2 which is focused on implementing numerous Quality of Service guarantees that will operate at the application level.