Humio Server 1.38.0 LTS (2022-03-15)

Version?Type?Release Date?Availability?End of Support

Security

Updates

Upgrades

From?

Config.

Changes?
1.38.0LTS2022-03-15

Cloud

2023-03-31No1.26.0Yes

Hide file hashes

Show file hashes

Download: https://repo.humio.com/repository/maven-releases/com/humio/server/1.38.0/server-1.38.0.tar.gz

Humio can now poll and ingest data from the Falcon platform's Falcon Data Replicator(FDR) service. This feature can be used as an alternative to the standalone fdr2humio project. See the Ingesting FDR Data into a Repository for more information.

New features and improvements

  • Falcon Data Replicator

    • Improved performance of FDRJob.

  • UI Changes

    • Minor UX improvements (ie. accessibility) on the queries panel.

    • On the time, bar and pie charts you can hold the ALT/OPTION key to display long legend titles.

    • When changing focus inside a dialog with the keyboard, the focus will no longer move outside the dialog while it is open.

    • Added a quick-fix for unknown escape sequences in the search field.

    • When using the table visualisation in dark mode, empty table cells are now clearly discernible.

    • First row entry in the statistics table on the repo page is now a table header and added hidden content to the empty table header in the new view page.

    • Added a warning for unknown escape sequences in the search field.

    • Hover information in the search field is shown despite an overlapping warning.

    • Reworked the hover message layout and changed the hover information on text (in the search field).

    • Better accessibility for queries panel. You can now tab to focus individual queries, and open a details panel. From here you can also access all actions in the details panel by tabbing.

    • Added a quick-fix to convert non-ASCII quotes to ASCII quotes in the search field.

    • Fixed a bug where the Package Marketplace would redirect to unsupported package versions on older Humio instances.

    • Hover over parameter names and arguments in the search field now includes the default value.

    • The Cluster Nodes table has been redesigned to allow for easier overview and copying the version-number.

    • Fixed an issue where queries with tail() would behave in an unexpected manner when an event is focused.

    • The bar and pie charts now support holding the SHIFT key to display unformatted numeric values.

    • Visually hidden clipboard field is now hidden for assistive technologies/keyboard users.

    • The search page now has focus states on the Language Syntax, Event List Widget and Save As buttons.

    • Pop-ups and drop-downs will now close automatically when focus leaves them.

  • GraphQL API

    • The PERMISSION_MODEL_MODE config option has been removed. All graphql related schema has also been removed.

    • Fixed a bug in the response from calling the installPackageFromZip GraphQL mutation. Previously, the response type exposed a deprecated clientmutationid that could not be selected. Also now if form fields are missing they are properly reported in the response.

    • Deprecates the ReadContents view action, in favor of ReadEvents. This also means ReadEvents has been undeprecated, as we have slightly changed how we consider read rights, and want the action names to match this.

  • Configuration

    • The Property inter.broker.protocol.version in kafka.properties now defaults to 2.4 if not specified. Users upgrading Kafka can either set inter.broker.protocol.version manually in kafka.properties, or pass DEFAULT_INTER_BROKER_PROTOCOL_VERSION as an environment variable to Docker when launching the container. Please follow Kafka's upgrade guidelines when upgrading a Kafka cluster to avoid data loss https://kafka.apache.org/documentation/#upgrade_3_1_0.

    • Reduce default value of INGESTQUEUE_COMPRESSION_LEVEL, the ingest queue compression level from 1 to 0. This reduces time spent compressing before inserting into the ingest queue by roughly 4x at the expense of a 10-20% increase in size required in Kafka for the ingest queue topic.

    • Added new configuration NATIVE_FALLOCATE_SUPPORT (default true) to allow turning off the use of fallocate and ftruncate internally.

    • Added config RDNS_DEFAULT_SERVER for specifying what DNS server is the default for the rdns query function.

    • Added new settings for how uploads to bucket storage are validated. In the case that validation with etags are not available, content length can be used instead.

    • When Kafka topic configuration is managed by Humio (default true) set max.message.bytes on the topics to the value of Config TOPIC_MAX_MESSAGE_BYTES, default is 8388608 (8 MB). Minimum value is 2 MB.

    • Added new configuration NATIVE_FADVICE_SUPPORT (default true) to allow turning off the use of fadvice internally.

    • Added config IP_FILTER_RDNS for specifying what IP addresses can be queried using the rdns query function.

    • Added config IP_FILTER_RDNS_SERVER for specifying what DNS servers can be allowed in the rdns() query function.

    • Added the config CORS_ALLOWED_ORIGINS a comma separated list for CORS allowed origins, default allows all origins.

    • Fixed a bug where TLS_KEYSTORE_TYPE and TLS_TRUSTSTORE_TYPE would only recognize lower-case values.

  • Functions

    • Fixed an issue where tail() could produce results inconsistent with other query functions, when used in a live query.

  • Other

    • Fixed an issue with epoch and offsets not always being stripped from segments.

    • Ensure only a cluster leader that still holds cluster leadership can force digesters to release partition leadership. This could cause spurious reboots in clusters where leadership was under contention.

    • For HTTP Event Collector (HEC) the input field sourcetype is now also stored in @sourcetype.

    • Published new versions of the Humio Kafka Docker containers for Kafka 3.1.0.

    • Added a new system-level permission that allows changing usernames of users.

    • During identity provider configuration, it's possible to fetch SAML configuration from an endpoint.

    • Improved off-heap memory handling. Humio now typically uses only 1 GB on systems with 32 vCPUs, down from typically 16 GB. This leaves more memory for other processes and page cache for data.

    • Fixed a compatibility issue with LogStash 7.16+ and 8.0.0 when using the Elasticsearch output plugin.

    • Improved the performance of deletes from global.

    • Do not run the Global snapshot consistency check on stateless ingest nodes.

    • Fixed an issue where users could be shown in-development feature on the client when running a local installation of Humio.

    • Fixed a bug in the Sankey chart such that it now updates on updated query results.

    • Added tombstoning to uploaded files, which helps with avoiding data loss.

    • Allow cluster managers access to settings for personal sandboxes and to block and kill queries in them.

    • Fixed an issue where top(max) could throw an exception when given values large enough to be represented as positive infinity.

    • Fixed an issue where live queries would sometimes double-count parts of the historic data.

    • Warn at startup if CORES > AvailableProcessorCount as seen by the JVM.

    • Fixed a bug where the Add Column button on the Fields panel would do nothing

    • Fixed an issue where queries of the form #someTagField != someValue ... would sometimes produce incorrect results.

    • Fixed a bug where providing a bad view/repository name when blocking queries would block the query in all views and repositories.

    • Fixed a compatibility issue with FileBeat 8.0.0.

    • Fixed several issues where users could add invalid query filters via the Add filter context button after selecting text in the Event List.

    • Fixed an ingest bug where under some circumstances we would reverse the order of events in a batch.

    • During Digest startup, abort fetching segments from other nodes if the assigned partition set changes while fetching.

    • Fixed an issue where negated functions could lose their negation.

    • Fixed an issue where percentile() would crash on inputs larger than ~1.76e308.

    • Previously a package could be updated with another package with the same name and version, but with different content. This is no longer allowed, and any attempt do so will be rejected and fail.

    • The Kafka client has been upgraded to 3.1.0 from 2.8.1. 3.1.0 enables the idempotent producer by default, which implies acks=all. Clusters that set acknowledgements to a different number via EXTRA_KAFKA_CONFIGS_FILE should update their config to also specify enable.idempotence=false.

    • LSP warnings don't crash queries any more.

    • Ensure a digester can only acquire or release partition leadership if no other digester has leadership of that partition. This could cause spurious reboots if digester leadership became contended.