RDS - Interval out of sequence

pmm server is connecting to RDS on aws env. it works well but, irregularly all rds cpu/select/locks is become increasing.
how can i fix this ?

/var/log/pmm-mysql-queries*.log

2016/10/27 19:31:00.001661 WARNING qan-analyzer-3c50e132 Skipping interval ‘4 2016-10-27 10:30:00 UTC to 2016-10-27 10:31:00 UTC (0-0)’ because interval ‘3 2016-10-27 10:29:00 UTC to 2016-10-27 10:30:00 UTC (0-0)’ is still being parsed
2016/10/27 19:31:00.013448 ERROR qan-analyzer-1ec9ef13-worker Got class twice: 86544a69f37dc3fd61efd03d8a0db40d
2016/10/27 19:32:00.001784 WARNING qan-analyzer-96c4d962 Skipping interval ‘40 2016-10-27 10:30:00 UTC to 2016-10-27 10:32:00 UTC (0-0)’ because interval ‘39 2016-10-27 10:28:00 UTC to 2016-10-27 10:30:00 UTC (0-0)’ is still being parsed
2016/10/27 19:32:00.001878 WARNING qan-analyzer-589032f7-worker Interval out of sequence: got 17, expected 16
2016/10/27 19:32:00.001929 WARNING qan-analyzer-3c50e132-worker Interval out of sequence: got 5, expected 4
2016/10/27 19:32:00.026592 ERROR qan-analyzer-1ec9ef13-worker Got class twice: 86544a69f37dc3fd61efd03d8a0db40d
2016/10/27 19:32:05.349868 WARNING data-sender Removed qan_1477564320543607456 because API returned 400: data.GetData: unexpected EOF
2016/10/27 19:33:00.001539 WARNING qan-analyzer-589032f7 Skipping interval ‘18 2016-10-27 10:32:00 UTC to 2016-10-27 10:33:00 UTC (0-0)’ because interval ‘17 2016-10-27 10:31:00 UTC to 2016-10-27 10:32:00 UTC (0-0)’ is still being parsed

Yes, we can expect some errors like that. If that’s not a problem then fine. We are constantly improving the tool and will have some fixes addressing mysql:queries services and perfschema query source on the next release as well.