Humio Server 1.29.0 Preview (2021-07-09)

VersionTypeRelease DateEnd of SupportUpgrades FromData MigrationConfig. Changes


This release has been revoked as it contained a known bug fixed in 1.29.1.

As a new feature Humio now includes an IOC (indicator of compromise) database from CrowdStrike to enable lookup of IP addresses, URLs and domains for malicious activity. This database is updated hourly. This is described in more detail at ioc:lookup()

Improvements, new features and functionality

  • GraphQL API

    • Removed the Usage feature flag which is now always enabled. This breaks backwards compatibility for internal graphql feature flag mutations and queries.

    • Field assignIngestToken was deprecated in Mutation type, use assignParserToIngestToken instead

    • Field createIngestListener was deprecated in Mutation type, use createIngestListenerV2 instead

    • Field copyParser was deprecated in Mutation type, use cloneParser instead

    • Field updateIngestListener was deprecated in Mutation type, use updateIngestListenerV2 instead

    • Removed the argument includeUsageView from the graphql mutation createOrganizationsViews which breaks backwards compatibility for this internal utility method.

    • Deprecated argument repositoryName was removed from Mutation.updateParser field

    • Deprecated argument name was removed from Mutation.updateParser field

    • Field addIngestToken was deprecated in Mutation type, use addIngestTokenV2 instead

  • Configuration

    • Humio nodes will now pick a UUID for themselves using the ZOOKEEPER_PREFIX_FOR_NODE_UUID prefix, even if Zookeeper is not used. This should make it easier to enable Zookeeper id management in existing clusters going forward.

    • Allow the internal profiler to be configured via an environment variable. See Environment Variables

    • Add a soft limit on the primary disk based on PRIMARY_STORAGE_PERCENTAGE and PRIMARY_STORAGE_MAX_FILL_PERCENTAGE (roughly the average of the two values). When the soft limit is hit and secondary storage is configured, the segment mover will prefer moving segments to secondary storage right away, instead of fetching them to primary and waiting for the secondary storage transfer job to move them.

  • Other

    • Return the responder's vhost in the metadata json.

    • Integrates the editing of alert searches and scheduled searches better with the search page

    • When editing a query, Enter no longer accepts a suggestion. Use Tab instead. The Enter key conflicted with the "Run" button for running the query.

    • Packages now support Webhook actions and references between these and alerts in the Alert schema.

    • Organization pages refactoring.

    • In thread dumps include the job and query names in separate fields rather than as part of the thread name.

    • Internal change to parsers adding an id, where previously they only had a name as key.

    • Enabled dark mode for cluster administration pages.

    • Speed up the SecondaryStorageTransferJob. The job will now delete primary copies much earlier after moving them to the secondary volume.

    • Scheduled searches are now allowed to run once every minute instead of only once every hour.

    • The "Save Search as Dashboard" Widget dialog now gives user feedback about missing input in a manner consistent with other forms.

    • Make GlobalConsistencyCheckerJob shut down more cleanly, it could previously log some ugly exceptions during shutdown.

    • Added dark mode support to Identity provider pages.

    • Previously, the server could report that a user was allowed to update parsers for a view, even though parsers cannot be used on views, only repositories. Now the server will always say the user cannot change parsers on views.

    • Created a new Dropdown component, and replaced some uses of the old component with the new.

    • Improved global snapshot selection in cases where a Kafka reset has been performed

Bug Fixes

  • Other

    • When testing a Parser and more events are returned in a test an info message is now displayed conveying that only the first event is shown.

    • Fixed an issue with "show in context" feature of the event list did not quote the field names in the produced query string.

    • Fixed an issue where exporting a saved query did not include the options for the visualization, e.g. column layout on the event list.

    • Fixed a bug in the blocklist which caused "exact pattern" query patterns to be interpreted as glob patterns.

    • Fixed race condition that could cause parsers to not update correctly in rare cases

    • Fixed an issue where the axis titles on the timechart were not showing up in dark mode

    • Fixed race condition that could cause event forwarding rules to not update correctly in rare cases

    • Fixed a bug causing match() to let an empty key field match a table with no rows.

    • Avoiding a costly corner case in some uses of glob-patterns.

    • Fixed bugs in the test parser UI, so that it should now always produce a result and be able to handle parsers that either drop events or produce multiple events per input event.

    • Fixed a bug where word wrapping in the event list was not always working for log messages with syntax highlighting (e.g. JSON or XML messages)

    • Fixed an issue related to validation of integer arguments. Large integer arguments would be silently truncated and lower limits weren't checked, which led to unspecified behavior. Range errors are now reported in the following functions:

    • Fixed a bug in the Search View. After editing and saving a saved query in the Search View, the notification message would disappear in an instant, making it impossible to read and to click the link therein.

    • Address edge cases where QueryScheduler could throw exceptions with messages similar to "Requirement failed on activeMapperCount=-36"