Skip to main content
After upgrading to a new platform version, always ensure that your installed component versions match the versions specified in the release notes. To verify this, navigate to FlowX.AI Designer → Platform Status.

​
Upgrade overview

FlowX.AI 5.9.2 is a maintenance release for the 5.9.x LTS line. It can be deployed as a direct upgrade from 5.9.1 with no breaking changes.

Upgrade Path

5.9.1 → 5.9.2: Direct upgrade with no breaking changes.Self-hosted customers still on the 5.1.x LTS upgrade directly to the 5.9.x line — see Migrating from 5.1 LTS.
Organization-ID migration fix (application-manager). 5.9.2 corrects the runtime-application organization-ID migration. If you worked around the original migration on 5.9.0 or 5.9.1 by manually marking it as executed, remove that record after upgrading to 5.9.2 so the corrected migration runs and converts your data. If you did not apply that workaround, no action is needed.
Manually marked Liquibase changesets. If any Liquibase changesets were manually marked as executed (MARK_RAN) to work around failures during earlier 5.9.x upgrade attempts, those records must be removed after upgrading to 5.9.2 so the corrected changesets can run. Contact FlowX support for the exact list of affected changesets before removing anything. If you did not mark any changesets manually, no action is needed.

​
Component versions

​
Core components

​
Embedded components

  • SpiceDB — see the 5.9.0 deployment guidelines for details.
  • Qdrant (vector database) — bundled in the AI Platform Helm chart; its version is managed by the chart and not pinned here. See the Qdrant setup guide for deployment and configuration.

​
Renderers

​
AI components

FlowX.AI groups AI services into four tiers by deployment role. AI Base components are the foundation; AI Runtime and AI Dev Time components both require Base to be installed first; AI Extensions are optional add-ons layered on top.
AI component versioning is unified in 5.9.2. All AI Platform services now ship under a single AI Platform version (10.9.0), including services that previously carried their own version lines (doc-parser, doc-converter, data-privacy, speech-to-text, web-crawler, evals-judge). Their 5.9.2 version numbers are therefore not comparable with earlier columns — for example, doc-parser moves from 10.10.1 to 10.9.0 without losing anything. Component rows below use the plain service names introduced in 5.9.2 (the ai-platform- prefix is dropped — see AI Platform service naming); earlier columns show the same component under its previous name.

​
AI Base components

​
AI Runtime components

​
AI Dev Time components

​
AI Extensions

For the plugins matrix, refer to the 5.9.0 deployment guidelines.

​
Kafka topics

5.9.2 introduces no new Kafka topics over 5.9.1. For the topics added in 5.9.1, see the 5.9.1 deployment guidelines; for the full multi-release topic matrix, see the 5.9.0 deployment guidelines.

​
Deployment notes

​
Standard deployment

This maintenance release follows standard FlowX.AI deployment procedures:
  1. Update container images to the 5.9.2 versions listed in the table above.
  2. Deploy services using your standard deployment process.
  3. Verify deployment via FlowX.AI Designer → Platform Status.

​
Configuration changes

  • Kafka reconnect backoff: platform services now default to a gentler broker-reconnect backoff — SPRING_KAFKA_PRODUCER_PROPERTIES_RECONNECT_BACKOFF_MS 1000 / SPRING_KAFKA_PRODUCER_PROPERTIES_RECONNECT_BACKOFF_MAX_MS 10000 (and the consumer equivalents) — and repetitive Kafka NetworkClient connection errors log at a quieter level, tunable via LOGGING_LEVEL_KAFKA_NETWORK_CLIENT (default ERROR). No action needed unless you previously tuned these.
  • Resource-server security type: the default for idp.resourceServer.securityType on the AI services changes to jwt. If you previously set the security type to jwt explicitly as a workaround, that override is no longer needed and can be removed.

​
Data partitioning and archiving

5.9.2 adds configuration for the archiving lifecycle on the process-engine and Integration Designer. All new settings default to off, so no action is required on upgrade.
  • Stuck-instance termination: an optional sweep (FLOWX_DATA_PARTITIONING_ARCHIVING_STUCKTERMINATE_ENABLED, default false) terminates instances that would otherwise block archiving. Enable it, plus its cron and batch-size, if archiving stalls on unfinished data.
  • Legacy backfill: one-shot backfills of partition_id (FLOWX_DATA_PARTITIONING_BACKFILL_ENABLED) and organization_id (FLOWX_DATA_PARTITIONING_ORGIDBACKFILL_ENABLED), both default false, populate legacy rows when you first turn partitioning on. The organization-ID backfill is optional on self-hosted and can remain false.
Partitions that still contain unfinished instances are not archived. Starting with 5.9.2, the archiving job detaches a partition only when every process instance in it has reached a finished state (FINISHED, FINISHED_WITH_ERROR, FAILED, EXPIRED, ABORTED, DISMISSED, or TERMINATED). A single instance still in CREATED, STARTED, or ON_HOLD blocks the entire partition; it is retried on the next archiving run. The same rule applies to workflow instances on the Integration Designer, where a STARTED workflow instance blocks its partition. If partitions accumulate unarchived, terminate the blocking instances — set process expiry, or turn on the stuck-instance termination sweep above.
For the full property tables and defaults, see Partitioning & archiving (process-engine) and Integration Designer setup → Workflow partitioning.

​
AI Platform service naming

AI Platform services are now deployed under their plain names — the ai-platform- prefix is dropped from pod, service, and image names (for example, ai-platform-agent-builder becomes agent-builder). If you have monitoring dashboards, alerts, or log queries that match the prefixed names, update them as part of this upgrade. The bundled Qdrant subchart is renamed the same way, so any explicit references to its service host must be updated too.
Core platform services resolve AI Platform components by these new names. Upgrade the AI Platform chart and the core platform services to 5.9.2 together — running a 5.9.1 AI Platform chart against 5.9.2 core services (or the reverse) breaks AI features.
Confirm any required configuration changes against the setup guides before upgrading.

​
Third-party supported component versions

Within each version line listed below, versions that have reached upstream end of life are not supported, even though their major/minor is listed. (*) Required only if using indexing via Kafka. Multiple sink connector implementations also work, as long as they are compatible with both the deployed Kafka and Elasticsearch version.
The table lists version lines, not individual versions. A version is supported for as long as its vendor still supports it, so check the version you plan to deploy against that vendor’s lifecycle page before committing to it. When a version reaches upstream end of life it is no longer supported; upgrade that component to keep your deployment on a supported configuration. Versions can also be added during the family’s lifetime as FlowX validates them.
For previous FlowX.AI releases, please refer to the Release Notes overview.
Last modified on July 30, 2026

Related topics

Deployment guidelines v5.6.0Deployment guidelines v5.3.0Deployment guidelines v5.4.0Deployment guidelines v5.5.0Deployment guidelines v5.2.0