Alfred Fuller
Alfred Fuller
Ideally, protoyaml (and protojson and protobuf) would only ignore fields explicitly marked as 'ignore' instead of all unknown fields. This would allow the safe removal of old fields without risking...
Thanks for the report! This protoyaml should treat this as an error.
Every change also requires a change to the json schema plugin. Protoyaml input has never been (and never needs to be) valid protojson. Only the output is normalized to be...
Honestly, because there are no backwards compatibility issues, there should be no option. This change only changes what used to be an error in to an intuitive non-error case. If...
Strong +1 from me as well. Wrapping causes a lot of extra, unneeded boilerplate code.
Sorry for the lack of detail. We've actually hit this particular limit in a number of ways in production. We are doing ~4k writes per second. These are very small...
"No matter how many operations are included in a TXN, they only increase the index & revision by 1." This is my point, why should we rewrite the application to...
(note that I suspect the workload we are running is significantly more write heavy than is typical, so just making the limit configurable would be enough)
FYI this is what we are using etcd for: https://buf.build/blog/bufstream-kafka-lower-cost and why our write rate is significantly higher than typical.