What's New in 6.6.0
New Features
Scale Out ClickHouse Cluster
In this release, you can set up a Scale Out ClickHouse Cluster using Supervisor and Worker nodes. All operations can be done from FortiSIEM GUI. The first step is to add storage to the FortiSIEM nodes for storing events. Then you configure the Cluster by specifying the number of shards and choosing the Keeper Cluster members and Server Cluster Replicas. Insertion, Replication and Queries are distributed to Supervisor and Worker nodes resulting in a redundant scale out and architecture.
For understanding concepts see Background Information.
For adding storage to Worker nodes, see Initial Configuration.
For configuring ClickHouse Cluster, see ClickHouse Configuration.
For configuring a ClickHouse retention policy, see Creating a ClickHouse Online Event Retention Policy.
For sizing guide on how to achieve high insert and query efficiency with ClickHouse, see the 6.6.0 Sizing Guide.
For ClickHouse Index Design and Query Optimization, see ClickHouse Usage Notes in the Appendix.
Generic REST API Based Event Collection
Increasingly Cloud Applications such as Cisco Umbrella, Microsoft Defender for Endpoint, WorkDay, Box.com, etc. are providing events and alerts via HTTPS based API. Rather than develop application by application support, this release provides a generalized way to configure an HTTPS based access method and pull data via that access method. A customer just needs to write a specific log parser for the application. This approach can cover all HTTPS based logging applications without requiring a new FortiSIEM release.
This release includes built in support for Cisco Umbrella Activity API and Microsoft Defender using Graph API using this Generalized HTTPS Access Method.
For details on configuring HTTPS API, see Generic Log API Poller HTTPS Advanced Integration from the External Systems Configuration Guide.
For details on creating HTTPS Credential for Cisco Umbrella Activity API, see Cisco Umbrella Configuration from the External Systems Configuration Guide.
For details on creating HTTPS Credential for Microsoft Defender API, see Windows Defender for Endpoint Configuration from the External Systems Configuration Guide.
Watchlist REST API for FortiGate Threat Feed and 3rd Party Devices
This list provides a simplified Watchlist Read REST API that FortiGate Firewalls or any other 3rd party device can use to download a FortiSIEM Watchlist and take blocking action. The REST API returns a list containing IP, Domain or Hash values that appear in FortiSIEM Watchlist.
Set Up
-
In FortiSIEM, set up a Watchlist and associate the Watchlist to one or more rules.
-
In FortiGate, define the Watchlist REST API as a threat feed.
-
See FortiGate/FortiOS Cookbook Threat feed for details on configuring FortiGate threat feed.
-
See Read APIs for Integration with FortiGate Firewalls for the FortiSIEM Watchlist Read API for FortiGates.
-
How it Works
-
A threat detection rule triggers in FortiSIEM and dynamically populates a Watchlist.
-
FortiGate periodically pulls the Watchlist as a threat feed.
-
FortiGate takes mitigating action.
Consider the scenario where the IP ages out from Watchlist in FortiSIEM because the rule stops triggering or other reasons.
-
FortiGate periodically pulls the Watch List again (that entry would be missing)
-
FortiGate undoes the mitigating action.
Key Enhancements
Custom Image Upload Endpoint
During Collector/Agent/Content upgrades, the Upgrade URL for Collectors and Agents is automatically generated by the App Server based on the Supervisor host name or IP in the GUI. However this approach does not work when there are Load Balancers in front of Supervisor node. This release provides an option for the user to specify a Load Balancer Host Name or IP and Supervisor will use it to create the custom endpoint for Collectors and Agents. If you use Load Balancer Host Name, it must be resolvable by Agents and Collectors. The Load Balancer is an easier choice.
For details on Setting up Custom Image Upload Endpoint, see Custom Update.
Performance Improvements
The following performance improvements have been made in 6.6.0.
Incident Handling Performance Improvement
App Server on Supervisor post-processes Incidents to add meta data, update risk scores, stores in PostGRESQL database and then executes notification policies and external integrations. FortiSIEM Manager aggregates Incidents from all Supervisors and stores them in local PostGRESQL database for display in FortiSIEM Manager GUI. Therefore, fast incident handling is critical for the system to work correctly at high loads.
This release contains extensive Incident handling performance optimizations. FortiSIEM Manager can handle about 1500 Incidents/sec from FortiSIEM Supervisors.
IdentityWorker Performance Improvement
This is achieved by making IdentityWorker multi-threaded while getting events from shared buffer.
ReportWorker and RuleWorker Performance Improvement
This is achieved by optimizing IN Group query handling code
QueryMaster Memory Usage Improvement
This is achieved by frequently calling memory release operations in Google tcmalloc library.
Update Glassfish CA Certificate store with Java CA Cert Store
Many Java based external integrations required users to import root CA certificates, as the Glassfish CA store was not populated after migrating to Glassfish V5 in an earlier FortiSIEM release. In this release, during migration and upgrade process, the Glassfish CA store is populated with valid certificates from the Java CA cert store. With this change, Java based external integrations should work more seamlessly.
System Upgrades
Upgrade to Rocky Linux 8.6 with patches released on May 16, 2022. See https://docs.rockylinux.org/release_notes/8_6/ and https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/8/html/8.6_release_notes/bug_fixes#bug-fix_securityfor more information.
Added Support for Elasticsearch 7.17
Support has been added for Elasticsearch 7.17.
New Device Support
-
Cisco Umbrella Activity API - done via Generalized HTTPS based Event Collection methodology
-
Microsoft Defender for end point using Graph API - done via Generalized HTTPS based Event Collection methodology
Bug Fixes and Minor Enhancements
Bug ID |
Severity |
Module |
Description |
---|---|---|---|
822029 |
Minor |
Anomaly |
Exclude network logon events from UEBA AI engine. |
824268 |
Minor |
App Server |
IPS to CVE validation does not work correctly. |
821804 |
Minor |
App Server |
Restarting App Server sometimes gives heap error and it is not fixed by re-deploying. |
816715 |
Minor |
App Server |
Org id in the device maintenance table did NOT get deleted when customer deletes the org. |
816492 |
Minor |
App Server |
Older open SAML library causes login failure to OKTA authentication portal. |
815696 |
Minor |
App Server |
When user changes org scope, create this event PH_AUDIT_USER_CHANGE_ORG_SCOPE instead of PH_AUDIT_USER_SUED. |
815025 |
Minor |
App Server |
Specifying Reporting device group and Event type Group does not work correctly for Event forwarding, Event Dropping and Retention Policy and Org mapping. |
813642 |
Minor |
App Server |
HTTP Error page reveals App Server name and version. |
811630 |
Minor |
App Server |
If you change the priority of a system rule, after FortiSIEM upgrade, the rule's priority is incorrectly reset to its default priority. |
821813 |
Minor |
Data |
Time Stamps are not parsed properly for Falcon Data Replicator. |
821585 |
Minor |
Data |
SystemShutdown event group has some incorrect event types. |
817295 |
Minor |
Data |
WinOSPull parser has incorrect function definition. |
814834 |
Minor |
Data |
Update BarracudaCloudGenFW parser format recognizer. |
814409 |
Minor |
Data |
Update FortiSIEM Claroty CTD parser. |
811907 |
Minor |
Data |
Reduce parser test events down to 10 for ease of cloning and testing. |
811224 |
Minor |
Discovery |
AWS credentials are included in aws-phgetflowlogs.php in clear text. |
821427 |
Minor |
ElasticSearch |
Add 7 day allowance to write data to Elasticsearch to cover data buffering more than 2 days. |
818548 |
Minor |
Event Pulling |
AWS Kinesis fails to sync shards and leases on connection. |
817081 |
Minor |
Event Pulling |
AWS Kinesis buffer size are incorrect and causes event pull failure. |
821263 |
Minor |
GUI |
Ticket Note not saved in Incident Ticket. |
820583 |
Minor |
GUI |
Show proper error codes returned by FortiSOAR executing playbook or connectors. |
814430 |
Minor |
GUI |
GUI does not allow dot in user name field. |
810866 |
Minor |
GUI |
When you disable a Device in Pull Events with a search filter, a different device in the global list is disabled. |
816499 |
Minor |
Parser |
Increase TCP/UDP buffer from 16KB to 24KB to handle large events up to 24KB. |
814318 |
Minor |
Report |
For ClickHouse Storage, Incident CSV export does not contain result for incident status and resolution. |
809386 |
Enhancement |
App Server |
The events PH_DEV_MON_LOG_DEVICE_DELAY_LOW and PH_DEV_MON_LOG_DEVICE_DELAY_HIGH do not generate reliably. |
802322 |
Enhancement |
App Server |
Changing Supervisor IP may lead to two Supervisor entries showing up in Cloud Health. |
801605 |
Enhancement |
App Server |
Users who have Read only admin permission on Super Global but have Full Admin on Org, cannot edit credentials in the Org after switching to that Org. |
804904 |
Enhancement |
App Server,GUI |
New Dashboard folders may not appear in the drop-down in the dashboard folder section. |
810382 |
Enhancement |
Data |
Some generic events from FortiGate need to be further parsed. |
809024 |
Enhancement |
Data |
Event Type generated from Windows logs received via Epilog is incorrect and very long. |
803091 |
Enhancement |
Data |
Update FortiGate parser to support Bandwidth Delta values. |
802247 |
Enhancement |
Data |
Update MSDefender AdvancedHunting parser. |
797026 |
Enhancement |
Data |
Update Nginx event parser. |
792333 |
Enhancement |
Data |
Update FortiAI parser to handle new log format and relabeling of product to FortiNDR. |
779162 |
Enhancement |
Data |
Update Claroty parser to accommodate year timestamp in syslog header and new event. |
774819 |
Enhancement |
Data |
Update FortiMail log parser. |
684254 |
Enhancement |
Data |
Update Extreme switch log parser. |
807886 |
Enhancement |
Data Manager |
A log from Elasticsearch event insert code contains password. |
801973 |
Enhancement |
Data Manager |
Online data is not removed when defining online retention policies for all events. |
807102 |
Enhancement |
Discovery |
SNMP V3 Trap support causes Auth Password and Priv Password directly in snmptrapd.conf on Collectors. |
796086 |
Enhancement |
Discovery |
Support more HP switches for discovery, performance and availability metrics via SNMP. |
775692 |
Enhancement |
Discovery |
Test connectivity results sometimes display Windows OMI text password in GUI. |
644096 |
Enhancement |
Discovery, Perf Monitoring |
Enable AES256 and SHA256 for SNMP V3. |
809387 |
Enhancement |
ElasticSearch |
Limit the number of dynamically allocated shards. |
795638 |
Enhancement |
Event Pulling |
Sophos logs via API are polled very frequently; thereby quickly reaching API limit. |
810548 |
Enhancement |
GUI |
System notifications are unreadable and overlaps with page information. |
783383 |
Enhancement |
Rule |
Suppress excessive PH_REPORT_PACK_FAILED events. |
Known Issues
-
Currently, Policy based retention for EventDB does not cover two event categories: (a) System events with phCustId = 0, e.g. a FortiSIEM External Integration Error, FortiSIEM process crash etc., and (b) Super/Global customer audit events with phCustId = 3, e.g. audit log generated from a Super/Global user running an adhoc query. These events are purged when disk usage reaches high watermark.
-
When retention policies are defined, there is a memory corruption issue in Parser module that can cause Parser module to crash or consume high memory. This may not always happen. This may result in event parsing being delayed, events being missed or Supervisor GUI being slow. From the Cloud Health page, you can see
phParser
process is down or has high CPU for any node. This issue has been resolved in release 6.6.2. -
On hardware appliances running FortiSIEM 6.6.0 or earlier, FortiSIEM
execute shutdown
CLI does not work correctly. Please use the Linuxshutdown
command instead. -
If you are running FortiSIEM 6.6.0 or earlier and have a Malware IOC group with 750K entries or more, then restarting the application server or rebooting of FortiSIEM Super node may fail. If this issue occurs, take the following steps:
-
ssh into the Supervisor as root user.
-
Run the following command to back up your copy of the domain.xml file.
cp -a /opt/glassfish/domains/domain1/config/domain.xml /opt/glassfish/domains/domain1/config/domain.xml.backup
-
Run the following command to modify the domain.xml file.
vi /opt/glassfish/domains/domain1/config/domain.xml
-
Find the following on line 220 and 221 (based off of default domain.xml file):
<jvm-options>-Xmx5120m</jvm-options>
<jvm-options>-Xms5120m</jvm-options>
and change to:
<jvm-options>-Xmx10G</jvm-options>
<jvm-options>-Xms10G</jvm-options>
-
Save the changes and exit.
-
Run the following command.
ps -ef | grep glassfish5
Example output:
admin 2993 1 85 Aug15 ? 1-16:43:25 /opt/Java/bin/java -cp /opt/glassfish5/glassfish/modules/glassfish.jar -XX:+UnlockDiagnosticVMOptions -XX:+UseParNewGC -XX:+CMSConcurrentMTEnabled -XX:+CMSParallelRemarkEnabled -XX:+LogVMOutput -XX:MaxPermSize=384m -XX:+UseConcMarkSweepGC -XX:ReservedCodeCacheSize=128m -XX:NewRatio=2 -XX:LogFile=/opt/glassfish5/glassfish/domains/domain1/logs/jvm.log -XX:+UseCompressedOops -XX:+DisableExplicitGC -XX:+UseCMSInitiatingOccupancyOnly -XX:CMSInitiatingOccupancyFraction=30 -Xmx5120m -Xms5120m -server -javaagent:/opt/glassfish5/glassfish/lib/monitor/flashlight-agent.jar -Djavax.net.ssl.trustStore=/opt/glassfish5/glassfish/domains/domain1/config/cacerts.jks -Dfelix.fileinstall.dir=/opt/glassfish5/glassfish/modules/autostart/ -Dcom.sun.aas.installRoot=/opt/glassfish5/glassfish -Dfelix.fileinstall.poll=5000 -Djava.endorsed.dirs=/opt/glassfish5/glassfish/modules/endorsed:/opt/glassfish5/glassfish/lib/endorsed -Djava.security.policy=/opt/glassfish5/glassfish/domains/domain1/config/server.policy -Dcom.sun.enterprise.server.logging.max_history_files=20 -Dosgi.shell.telnet.maxconn=1 -Dfelix.fileinstall.bundles.startTransient=true -Dcom.sun.enterprise.config.config_environment_factory_class=com.sun.enterprise.config.serverbeans.AppserverConfigEnvironmentFactory -Dfelix.fileinstall.log.level=2 -Djavax.net.ssl.keyStore=/opt/glassfish5/glassfish/domains/domain1/config/keystore.jks -Djava.security.auth.login.config=/opt/glassfish5/glassfish/domains/domain1/config/login.conf -Dfelix.fileinstall.disableConfigSave=false -Dorg.owasp.esapi.resources=/opt/phoenix/config -Dfelix.fileinstall.bundles.new.start=true -Dcom.sun.aas.instanceRoot=/opt/glassfish5/glassfish/domains/domain1 -Dosgi.shell.telnet.port=37401 -Dgosh.args=--nointeractive -Dcom.sun.enterprise.security.httpsOutboundKeyAlias=s1as -Dosgi.shell.telnet.ip=127.0.0.1 -DANTLR_USE_DIRECT_CLASS_LOADING=true -Djava.awt.headless=true -Djava.ext.dirs=/opt/Java/lib/ext:/opt/Java/jre/lib/ext:/opt/glassfish5/glassfish/domains/domain1/lib/ext -Djdbc.drivers=org.apache.derby.jdbc.ClientDriver -Djava.library.path=/opt/glassfish5/glassfish/lib:/opt/phoenix/lib64:/opt/phoenix/lib32:/usr/local/lib:/usr/java/packages/lib/amd64:/usr/lib64:/lib64:/lib:/usr/lib com.sun.enterprise.glassfish.bootstrap.ASMain -upgrade false -domaindir /opt/glassfish5/glassfish/domains/domain1 -read-stdin true -asadmin-args --host,,,localhost,,,--port,,,4848,,,--secure=false,,,--terse=false,,,--echo=false,,,--interactive=false,,,start-domain,,,--verbose=false,,,--watchdog=false,,,--debug=false,,,--domaindir,,,/opt/glassfish5/glassfish/domains,,,domain1 -domainname domain1 -instancename server -type DAS -verbose false -asadmin-classpath /opt/glassfish/lib/client/appserver-cli.jar -debug false -asadmin-classname com.sun.enterprise.admin.cli.AdminMain
Note: The Process id is 2993 from the example.
-
From step f, note the Process id (PID) and kill it, by running the following command, and substituting <PID> with your actual PID.
kill -9 <PID>
Example command:
kill -9 2993
-
Wait for
AppServer
to restart.
-
-
FortiSIEM 6.5.0 ran ClickHouse on a single node and used the Merge Tree engine. FortiSIEM 6.6.0 onwards runs Replicated Merge Tree engine, even if Replication is not turned on. So after upgrading to FortiSIEM 6.6.0, you will need to do the following steps to migrate the event data previously stored in Merge Tree to Replicated Merge Tree. Without these steps, old events in 6.5.0 will not be searchable in 6.6.0. Once you are on post 6.5.0 release, you will not need to do this procedure again.
To upgrade your FortiSIEM from 6.5.0 to 6.6.0 or later, take the following steps.
-
Navigate to ADMIN >Settings > Database > ClickHouse Config.
-
Click Test, then click Deploy to enable the ClickHouse Keeper service which is new in 6.6.0.
-
Migrate the event data in 6.5.0 to 6.6.0 by running the script
/opt/phoenix/phscripts/clickhouse/clickhouse-migrate-650.sh
.
-
-
Sometimes App Server may not come up properly after upgrading to 6.6.0. This is rare and Fortinet has only seen it occur in Azure, although the hypervisor platform has very little to do with it. If this issue occurs, you will see that the backend ph processes will be down and there will not be any upgrade errors in ansible. In this situation, use the following workaround to get the system up and running.
-
Download
gf_admin-keyfile
anddeploy-fresh.sh
. -
Copy the file
gf_admin-keyfile
to Fortisiem Supervisor node under/opt/phoenix/deployment/jumpbox
. -
Login to FortiSIEM Super console and run the following command.
cp /opt/phoenix/deployment/deploy-fresh.sh /opt/phoenix/deployment/deploy-fresh.sh.orig
-
Copy the file
deploy-fresh.sh
to FortiSIEM Super under/opt/phoenix/deployment
. -
Login to the Supervisor console and run the following commands.
chmod +x /opt/phoenix/deployment/deploy-fresh.sh
su - admin
/opt/phoenix/deployment/deploy-fresh.sh /opt/phoenix/deployment/phoenix.ear
-
-
If you are running ClickHouse and upgrade from 6.5.0 to 6.6.0 and go to Storage > Online Settings and click Test, it will fail. Fortinet introduced a new disk attribute called "Mounted On" to facilitate disk addition/deletion that was not present in 6.5.0. Follow these steps to fix the problem.
-
Go to ADMIN > Setup > Storage > Online. ClickHouse should be the selected database.
-
For Hot tier and for every configured disk within the tier, do the following:
-
The existing disk should have empty Mounted On.
-
Click + to add a disk. For the new disk, Disk Path should be empty and Mounted On set to /data-clickhouse-hot-1.
-
Copy the Disk Path from the existing disk into this newly disk. The new disk should have the proper Disk Path and Mounted On fields.
-
Delete the first disk with empty Mounted On.
Do this for all disks you have configured in 6.5.0. After your changes, the disks should be ordered /data-clickhouse-hot-1, /data-clickhouse-hot-2, /data-clickhouse-hot-3 from top to bottom.
-
-
Repeat the same steps for the Warm tier (if one was configured in 6.5.0), except that the Mounted On fields should be /data-clickhouse-warm-1, /data-clickhouse-warm-2, /data-clickhouse-warm-3 from top to bottom.
-
When done, click Test, then click Deploy.
-
-
In Elasticsearch based deployments, queries containing "IN Group X" are handled using Elastic Terms Query. By default, the maximum number of terms that can be used in a Terms Query is set to 65,536. If a Group contains more than 65,536 entries, the query will fail.
The workaround is to change the “max_terms_count” setting for each event index. Fortinet has tested up to 1 million entries. The query response time will be proportional to the size of the group.
Case 1. For already existing indices, issue the REST API call to update the setting
PUT fortisiem-event-*/_settings { "index" : { "max_terms_count" : "1000000" } }
Case 2. For new indices that are going to be created in the future, update fortisiem-event-template so those new indices will have a higher max_terms_count setting
-
cd /opt/phoenix/config/elastic/7.7
-
Add
"index.max_terms_count": 1000000
(including quotations) to the “settings” section of thefortisiem-event-template
.Example:
...
"settings": { "index.max_terms_count": 1000000,
...
-
Navigate to ADMIN > Storage > Online and perform Test and Deploy.
-
Test new indices have the updated terms limit by executing the following simple REST API call.
GET fortisiem-event-*/_settings
-
-
If you set up Disaster Recovery (DR) on FortiSIEM 6.5.0, then upgrading the Secondary to 6.6.0 will fail. If you set up DR on older versions, this issue does not occur. The following workarounds are available, depending on your situation.
If you have not started the upgrade to 6.6.0 on DR yet, take the following steps:
Instructions before Upgrading Secondary in DR Configuration
-
Step 1: Back up Primary Glassfish Key into Secondary
-
Step 2: Insert into Secondary before Upgrade
-
Log into the Secondary Supervisor as root.
-
Run
phsecondary2primary
Note: This disables disaster recovery for the time being.
-
Run the following command to modify the
admin-keyfile
file.vi /opt/glassfish/domains/domain1/config/admin-keyfile
-
Paste the output from Step 1 v. to the
admin-keyfile
file, replacing the current entry and save.
-
-
Step 3: Verify that the Key has Taken Effect on Secondary Supervisor
-
Run the following commands.
su - admin; /opt/glassfish/bin/asadmin stop-domain /opt/glassfish/bin/asadmin start-domain /opt/glassfish/bin/asadmin login
For user name, enter admin
For password, enter the password from Step 1 iii.
Example success login:
/opt/glassfish/bin/asadmin login Enter admin user name [Enter to accept default]> admin Enter admin password> Login information relevant to admin user name [admin] for host [localhost] and admin port [4848] stored at [/opt/phoenix/bin/.gfclient/pass] successfully. Make sure that this file remains protected. Information stored in this file will be used by administration commands to manage associated domain. Command login executed successfully.
Note: Step 3 must work in order to proceed to Step 4.
-
-
Step 4: Upgrade to 6.6.0 on Secondary
-
Follow the Upgrade Guide.
-
-
Step 5: Add Secondary Back into the System
-
Log into the Primary Supervisor’s GUI.
-
Navigate to ADMIN > License > Nodes.
-
Select the Secondary entry in the GUI and click on Edit.
-
Click on Save to re-establish connection to the Secondary.
-
Instructions if you are Already in a Failed State in the Secondary
-
Step 1: Grab the Supervisor's DB Password
-
Step 2: Update the Glassfish Password on Secondary to Continue with the Upgrade
-
Log into the Secondary Supervisor as root.
-
Run the following commands to modify the
admin-keyfile
file.cd /opt/glassfish/domains/domain1/config/
cp -a admin-keyfile admin-keyfile.bad
vi /opt/glassfish/domains/domain1/config/admin-keyfile
-
Paste the output from Step 1 v. to the
admin-keyfile
, replacing the current entry and save.
-
-
Step 3: Deploy
appserver
Manually after Admin Passkey has Changed-
Run the following command.
vi /opt/phoenix/deployment/deploy-fresh.sh
-
Replace the password with the password from Step 1 iii by finding the following line, and making the password replacement there.
echo "AS_ADMIN_PASSWORD="$dbpasswd > $DEPLOYMENR_HOME/glassfish-pwd.txt
Example:
echo ‘AS_ADMIN_PASSWORD=ad1$dnsk%’ > $DEPLOYMENR_HOME/glassfish-pwd.txt
Note: The use of single quote character (‘) in the replacement vs the double quote character (“).
-
Save the file
deploy-fresh.sh
. -
Run the following commands.
su - admin
/opt/phoenix/deployment/deploy-fresh.sh /opt/phoenix/deployment/phoenix.ear
-
-
Step 4: Modify Ansible Playbook to Finish Upgrade
-
Run the following command.
vi /usr/local/upgrade/post-upgrade.yml
-
Remove only the following:
- configure - update_configs - fortiinsight-integration - setup-python - setup-node - setup-clickhouse - setup-zookeeper - setup-redis - migrate-database - appserver
-
Save the modification.
-
Resume the upgrade by running the following command.
ansible-playbook /usr/local/upgrade/post-upgrade.yml | tee -a /usr/local/upgrade/logs/ansible_upgrade_continued.log
-
Reboot the Supervisor if system does not reboot itself.
-
-
-
FortiSIEM uses dynamic mapping for Keyword fields to save Cluster state. Elasticsearch needs to encounter some events containing these fields before it can determine their type. For this reason, queries containing
group by
on any of these fields will fail if Elasticsearch has not seen any event containing these fields. Workaround is to first run a non-group by query with these fields to make sure that these fields have non-null haves.