Hi all,
Backend Infra team has landed the changes to flush online-monitor metrics regularly. Previously, it only flushes the metrics when there is a datapoint at newer period || when the online monitor is shut down. There was a problem for metrics which datapoints are very sparse (for example: exception count), because the timing of the metrics submission could be off for hours or even days. Now, it will try to flush the metrics every 1 minute (default flush interval).
Thank you!
Note: the related diff is https://phabricator.noc.tvlk.cloud/D80681