> Thanks Adrian for taking the time to respond. I will review the > documentation once more, just in case I missed anything. > > My apologies - I forgot to mention in my original post, that our last > cluster shutdown was over 15 days ago, thus an "unclean" > shutdown would not explain these daily stats reset that we are seeing. > It might also be relevant to mentioned
Then something is resetting the statistics.
1) What is log_min_error_statement set to?
2) Did you reload the server when changing?:
log_min_duration_statement | 0 log_statement | all