On-Premise Reliability Improvements
Improvement
The on-premise AlphaSOC runtime has received a set of reliability improvements focused on upgrades and large data volumes.
- Resilient upgrades: Sigma rules that fail validation during an upgrade are now logged as warnings and skipped, so a single invalid rule no longer blocks the runtime from starting. You can fix flagged rules on your own schedule.
- Safe upgrades with queued data: the internal queue records the format version of each stored payload, so an upgraded runtime reads both the previous and the current format and drains anything left from the older version instead of stranding it. Future format changes roll out the same way, without discarding data that is still queued.
- Safer internal migrations: internal queue migrations now run in resumable batches, and the free disk space check is sized against the data actually being moved rather than a fixed worst case, so upgrades on smaller deployments no longer ask for headroom they do not need.
- Lower memory usage: large imports are now streamed file by file rather than buffered in batches, keeping memory usage stable on high-volume deployments.