It can take seconds (at least locally) to write a few thousand datapoints using one of the store_*_values procedures. This is also probably leading to a ~10 reads for every row written. Now all the values are bundled into a json array instead of going through each individual value with at least a 10x to 100x speedup.
It can take seconds (at least locally) to write a few thousand datapoints using one of the store_*_values procedures. This is also probably leading to a ~10 reads for every row written. Now all the values are bundled into a json array instead of going through each individual value with at least a 10x to 100x speedup.