Changelog for version 5.4.6#
Released 15 March 2024#
Gravwell#
Additions#
Added a button in Query Studio to Apply a timeframe without launching a search.
Added the ability for Actionable readers to access the form page with disabled inputs.
Added the ability to enter multi-line Secrets.
Added a retry for a failed attempt to pull results for an Alert.
Added the ability to share write access with a group for Scheduled Searches, Flows, and Alerts.
Any Scheduled Search (dispatcher) or Flow (consumer) that you have access to can be added to an Alert - even if you do not own the Scheduled Search or Flow.
Bug Fixes#
Fixed an issue where ingest would fail and retry with overly dramatic logs when attempting to write to a block that was actively aging out.
Fixed an issue where a search far into the future would consume significant CPU on the webserver in a cluster environment.
Fixed an issue where a “beginning of line” regex delimiter could cause a dropped buffer while waiting for next delimiter and potentially cause data loss in File Follower.
Fixed an issue where a user could see cached webpages using an expired license and the browser Back button.
Fixed an issue where uploading a kit could show a duplicate in a different state.
Fixed an issue where a Gravwell API tokens were not respected when hitting an Alerts endpoint.
Fixed an issue with writing back to files when performing searches that caused stress on COW file systems.
Fixed an issue with failover well feeder locking when aborting queries.
Fixed an incorrect type assertion that could cause a crash in the
slice
module.Fixed an issue with indexer shutdown related to timeouts in network connectivity.
Fixed an issue with detecting and handling oversized blocks in the ingest server.
Fixed an issue with bounds checking in the ipfix packet parser.
Fixed an issue with tile metadata in Dashboards.
Fixed an issue with creating a Scheduled Search from Query Studio when using a custom duration timeframe.
Fixed an issue with performance on the Persistent Searches page when there are a large number of searches.
Improved problems with extremely long launch delays when replication was backed by very low IOP storage.
Improved the way the webserver shuts down.
Made IP-based filters with no CIDR notation imply a /32 or /128.
Ingesters#
Updated HTTP ingester to use AWS Firehose naming schemes.
Additions#
For a timestamp that is zero or some very low value, the HEC ingester will now use the ingest time instead.
Bug Fixes#
Fixed an issue with the HTTP ingester running out of memory upon mass reconnect or failure to ingest.
Improved Federator throughput when lots of indexers are present.