- ISSUE=1971 Fixed edge case where in some cases queries to TSDS services would fail but that failure would not be properly trapped and requeued, resulting in missed aggregations.
- ISSUE=530 Added knob to control maximum amount of data fetched in a single block. This allows better control over server behavior and should help with situations like data backfill from choking up the server.
- ISSUE=821 Slight change in architecture to fork a new temporary process that does the actual work to work around perl memory management issues. Impact should be that while idling until next run considerably less memory should be held onto.
- ISSUE=708 Fixed edge case where writer would complain about aggregate messages with a histogram width with a floating point value. This was only seen in cases where data variances were very small, such as in optical readings.
- ISSUE=13132 Fixed issue where upon first install the daemon would make the determination that everything needed aggregation.
- ISSUE=12464 Overhaul of TSDS data aggregation process. Instead of being bundled with the services library, this is now a standalone package to better enable horizontal distribution. Data retrieval now happens through the Query interface to enable aggregates from aggregates, and detection of what needs to be aggregated is more intelligent.