In reference to issue #2, question was asked by @Khomutov-SY regarding meta data at BGS. @SimonFlower replied with:
At the Edinburgh GIN we store data in a database, not in any of the formats that the data is sent to us. This has a number of advantages: it is easy for us to handle any incoming data format; we can easily cope with fragments of day files and duplication of data, when a new format is introduced (such as Intermagnet CDF) we can make the format available for all data very quickly. The disadvantage is that we don't store the metadata from any of the formats.
It has been on my mind for some time to start a GIN metadata database alongside the data database, to collect the sort of metadata that Sergey is talking about. I know that many observatories put effort into their metadata and it is a shame that the Edinburgh GIN loses this metadata. I will put some thought into this issue (it also affects the CDF format, although I don't think anyone is delivering data to the GIN in CDF).
I don't think the other GINs work in this way. I think that metadata is preserved when sending data through them?
For Canada, as the current endpoint, we don't alter the products however we also don't create other dynamic products like BGS does and is very useful (ex: CovJSON). Having one format to generate others adds simplicity and is very nice perk offered by BGS. Finding a storage format that very flexible is the best solution but its understandable that its not easy to do and requires thinking (like you mentioned @SimonFlower).
In reference to issue #2, question was asked by @Khomutov-SY regarding meta data at BGS. @SimonFlower replied with:
For Canada, as the current endpoint, we don't alter the products however we also don't create other dynamic products like BGS does and is very useful (ex: CovJSON). Having one format to generate others adds simplicity and is very nice perk offered by BGS. Finding a storage format that very flexible is the best solution but its understandable that its not easy to do and requires thinking (like you mentioned @SimonFlower).