We have completed our efforts to stem the growing backlogs in the aggregation layer by shifting more resources there without destabilizing the layer and the storage. We're now tuning some parameters to speed up the consumption of backlogs.
Posted Dec 15, 2019 - 04:11 UTC
Due to intermittent data center network issues, there are backlogs of data points that have not been pushed into permanent storage. As a result, gaps may appear in graphs when viewing 1-10 hours or greater than 5 days. No data points have been dropped but data point backlogs need to be replayed. Ingestion has not been impacted and other resolutions should be displaying as expected. We're working to carefully shift resources so as not to overload the storage in order to resolve the backlog as quickly as possible.