Migration strategy going forward

Hey all

It seems that pushing aws-sdk-cpp to 1.9.x (after being stuck on 1.8 for ages) did not cause the world to go up in flames for [arrow](https://github.com/conda-forge/arrow-cpp-feedstock/issues/918)[^1], which is great. 🥳

[^1]: effectively the only ABI-affected consumer in conda-forge

But now the world has moved on to 1.10, and upstream arrow is also [thinking](https://github.com/apache/arrow/issues/20272#issuecomment-1380120591) about upgrading to that.

Since pyarrow feedstock CI is extremely costly (4 maintenance branches which need several manual restarts due to timeouts), we cannot migrate this feedstock all the time, there are _way_ too many versions for that.

But I think we could do something every 20, 25 or 50 patch versions. I'd like to then create a respective abi_migration_branch for that version here (like I did for [1.9](https://github.com/conda-forge/aws-sdk-cpp-feedstock/tree/v1.9.x)), so that this version will keep getting updates from the rest of the aws-c-* ecosystem while we hold it stable for a while.

Since we've now reached aws-sdk-cpp 1.10.50, I thought it's perhaps a good time to discuss this.

CC @conda-forge/aws-sdk-cpp @conda-forge/arrow-cpp

Also tagging @conda-forge/core, because the pinning repo is full of PRs for aws_sdk_cpp versions bumps, and I guess we should agree on strategy / frequency.