Skip to main content

Version 1.17.0

Release Date: July 25, 2026 Release Type: Release Candidate

This is a Release Candidate (pre-release). After validation the final 1.17.0 release will be published. The headline of this release is the XI connectivity suite — native XI 3.0 messaging, an ESR repository facade, and a unified Connections workspace — together with an ActiveMQ connector, resilient monitoring storage, and per-flow log retention.

Backend Server​

New Features​

  • XI 3.0 inbound connectivity: MIP can receive XI 3.0 messages directly from partner registry/business systems through an always-on inbound listener on a configurable public path (default /XIMIP). It parses XI 3.0 envelopes, resolves the sending and receiving parties, routes each message to the correct flow, and replies with standards-compliant acknowledgements and faults. Error handling is fail-closed, so a message is never falsely acknowledged, and rejections (unknown binding, malformed envelope, authentication failure) now appear in Message Monitoring with an actionable explanation.
  • Reliable asynchronous XI messaging (EO / EOIO): Asynchronous XI messages are accepted into a durable inbox and persisted before being acknowledged, so no message is lost under failure. Exactly-Once deduplicates repeats and Exactly-Once-In-Order keeps strict per-queue ordering, with safe handling across multiple server nodes, automatic cleanup of old records, and flow-declared ordered-queue names. A background worker sends acknowledgements back to the sending system using pooled credentials, and each message is linked to its Message Monitoring record.
  • XI 3.0 outbound requests: A new flow step sends XI 3.0 requests to a registry/business system selected by name, supporting Best Effort and Exactly Once quality of service with automatic retry on transient failures, and returns the peer's response into the flow.
  • ESR repository facade: MIP can directly answer a partner system's repository (ESR) browser requests, so the interfaces deployed in MIP flows become discoverable from the design tooling without maintaining a separate repository. Includes a capture/hybrid mode to ease migration, a query and public-WSDL service, and authentication of incoming repository requests against MIP user credentials.
  • RFC function browsing and automatic schema generation: When configuring an RFC connector, users can test the connection, browse the remote function modules on the target system, and automatically generate request and response schemas from a function's interface, removing manual schema authoring.
  • ActiveMQ connector: New connector to send to and receive from external ActiveMQ brokers, supporting both OpenWire and AMQP protocols with TLS and username/password authentication.
  • Kafka commit strategy and advanced consumer options: Kafka consumer flows gain an "after flow" commit strategy that advances the consumer offset only once the flow finishes processing successfully (with bounded, validated retry), preventing message loss on failures or restarts. A set of optional advanced consumer tuning fields (offset reset, poll interval, fetch sizes, isolation level, manual/auto commit) is also exposed, each applied only when set so existing flows are unaffected.
  • Kafka topic administration: A new capability lists topics with their status, shows per-topic configuration and per-consumer-group lag, and lets operators change editable settings (retention, cleanup policy, compression, maximum message size) and increase partition counts, validating every change before it reaches the broker.
  • Automatic monitoring failover between Elasticsearch and PostgreSQL: When Elasticsearch is the monitoring store and becomes unreachable or unable to accept writes, message-monitoring records automatically fail over to PostgreSQL and fail back on recovery, so no monitoring data is lost during an Elasticsearch outage. It detects both connectivity loss and write blocks such as a full disk, a red cluster, or rejected writes. Opt-in.
  • Side-effect-free Flow Simulation: Dry-run and mock simulations are now guaranteed not to touch external systems or pollute monitoring — they no longer write files, upload over FTP/SFTP, send e-mail, produce to Kafka/JMS/Solace, call cloud services, invoke RFC/OFTP2/WebDAV/MCP or external databases, and create no monitoring, attachment or trace records. Live Test still runs for real and is recorded, exactly as before.
  • REST multipart / form-data handling: The inbound REST listener now correctly preserves and forwards multipart/form-data request bodies, and can optionally record each part in the attachment log, with a configurable size cap to protect memory against very large uploads.
  • Per-flow log retention and longer default: Log retention can be configured per flow, so a flow can keep its logs longer or shorter than the global default. The default log retention has been raised from 5 to 14 days.
  • "No Payload" logging level: A new logging level records message counts and success/failure without storing payload bodies on success, while still capturing payloads on error, reducing stored data volume and helping with data-privacy requirements.
  • OpenShift deployment support: Official OpenShift-native deployment manifests (restricted security context, an operator-managed PostgreSQL, native Routes) plus additional rootless container image variants, alongside the existing Kubernetes/Azure option.
  • OFTP2 delivery-error visibility: OFTP2 transfer problems (a duplicate-file rejection, or a partner disconnecting without a delivery receipt) now surface in Message Monitoring as clear errors, and virtual-file identifiers are made unique to avoid false duplicate-file rejections on resend.
  • Service-user name in authenticated logs: For Basic- and JWT-authenticated inbound requests, message logs record the authenticating service user's name, giving clearer traceability of who sent a message.

Bug Fixes​

  • Binary request bodies (for example an image or PDF uploaded as a multipart part) now log correctly; previously embedded null bytes broke log persistence and the corresponding attachments appeared missing in monitoring.
  • The Set Context node now produces valid JSON when a mapped value is null or an empty array/object, instead of malformed output that broke downstream processing.
  • Attachment logging (when Elasticsearch is disabled) now uses parameterized database writes, closing a SQL-injection exposure, and stamps proper timestamps so retention and time-window filtering work.
  • XI receiver faults now carry the real error category, code and detail into the flow error and monitoring, so operators can distinguish an interface mismatch from a missing receiver.
  • Basic-authentication credential lookup for the XI connector now resolves correctly instead of always returning empty.
  • A flow's per-flow global-configuration override now only blocks an inherited setting; it can no longer force-enable a setting or replace a global's value.
  • The Service Methods (XSD/WSDL) generator now includes operation descriptions and honors a configured target namespace.
  • On Kubernetes, the bundled Kafka broker now retains its topics, consumer offsets and metadata across pod restarts.

Frontend Server​

New Features​

  • Connections suite — XI/PO, RFC/IDoc and web services: Connect to a PO/XI system, browse its software components, and import ESR service interfaces with cached WSDLs; browse RFC destinations and import function interfaces, materializing their import/export schemas as flow resources; and discover web services through a WS-Inspection (WSIL) catalog and import selected services. Register XI business systems, monitor XI inbound queues with retry and cancel, and use the new XI proxy receiver step.
  • ActiveMQ connector: Server-side support for building flows that send to and receive from external ActiveMQ brokers.
  • EDI converter connector and EDI schema management: A unified connector for converting EDI messages, plus management of the EDI schemas your flows rely on.
  • Kafka topic browser and advanced consumer settings: Browse the Kafka topics used by your flows or across the cluster — partitions, health status, windowed producer/consumer counts and consumer-group lag — change topic settings, and refresh live. Fine-tune Kafka consumers on the start connector (offset reset, poll/fetch sizing, isolation level).
  • Flow simulation with a reusable mock library: Run a simulation before deploying a flow, watch step-by-step results stream in, keep a history of past runs, and save reusable mock responses to re-apply without retyping them.
  • Guided import with conflict analysis: Import a package or flow-mapping archive with an up-front analysis of exactly which flows, resources and mappings already exist, then choose overwrite or skip per item or in bulk before applying. It runs in the background with progress and cancellation, gives clear messages for empty or invalid archives, and leaves the existing one-click import unchanged.
  • Global flow configuration with per-flow overrides: Define reusable configuration values once and apply them across flows, or override/block them for individual flows. A flow's values now travel with it on export and import, and deletion is guarded by a where-used check.
  • Detailed flow change history: Each flow update records exactly what changed — which steps were added, removed or modified and which fields changed — filtered to the fields actually visible in the designer.
  • Per-flow log retention, historical counts and longer default: Set a custom retention window for individual flows; report successful and error message counts per flow over any date range even after logs are cleaned up; and the default retention is now 14 days instead of 5.
  • Message log identity tracking and export: Message logs capture who triggered each message and which credential and identity were used, and these fields are included in the CSV export.
  • Resilient, timezone-correct monitoring: If Elasticsearch becomes unavailable, monitoring automatically reads from the database and switches back on recovery, re-syncing the missed window. The time-range filter interprets dates in your configured business timezone and honors an explicit offset, and suppressed per-step payloads stay hidden in Show Full Log.
  • REST multipart/form-data support: The REST start connector can optionally accept inbound multipart/form-data uploads, log the attachments, and forward the body.
  • OAuth2 signed-JWT client authentication and login improvements: When MIP requests OAuth2 tokens, it can authenticate with a signed JWT (client_secret_jwt or private_key_jwt) instead of a shared secret, selectable per credential. OAuth2 login redirects resolve the correct public address, the interface knows when OAuth2 login is available, and an unreachable gateway is handled gracefully.
  • Single-credential mode and improved service-user management: An optional mode where each connection carries one automatically named credential created atomically with the connection, plus a paginated, sortable, searchable two-panel service-user view scoped by role.
  • Connector refinements: X.400 connector settings were refined around O/R addressing, TLS, delivery reports and body-part type (existing flows continue to load); FTP and SFTP connections accept additional custom parameters; and OFTP2 connections show which flows use them on both sending and receiving sides, blocking deletion while in use.
  • OpenShift support and health probes: Native OpenShift deployment manifests and rootless container images, plus liveness and readiness probes that fail fast when the database is unreachable, for reliable rolling deployments.

Bug Fixes​

  • Version-history activity entries are stamped with their own save time instead of the previous save's time.
  • The credential Confirm Sync action now creates the API management consumer when one does not yet exist, so the credential reaches the gateway and appears in the Consumer list.
  • Deleting a service user cleans up its related OAuth2 and API-management bindings, avoiding orphaned records.
  • Certificate Where Used now includes Kafka connectors, so certificates in use by Kafka flows are listed and protected from deletion.
  • The AI-monitoring certificate table lists only certificates within the expiry threshold or already expired, matching its banner.
  • Importing a flow whose name collides with an existing flow no longer fails — it is renamed automatically.
  • API management route provisioning retries while the gateway settles, and MIP no longer connects to the gateway at all when API management is turned off.

Frontend Web​

New Features​

  • Design flows directly from XI (ESR) interfaces: New inbound and outbound XI Proxy connectors let you build flows straight from interfaces in the Enterprise Services Repository — pick an interface, target a registered system, choose its WSDL source, bind sender/receiver party-service, and set EOIO queues, all from a searchable catalog tree grouped by software-component version and namespace.
  • XI systems registry and queue monitoring: A new administration area to register XI/PO systems, manage a shared credential pool and business-party bindings, copy an existing system definition, run detailed connection tests, and monitor message queues.
  • Import XI interfaces with progress and where-used: Import interfaces with multi-select of software components to scope large catalogs, preview and download pretty-printed WSDL, follow live import progress, and see exactly which flows use an imported interface before you change it.
  • RFC connectivity — browse, explore, test and schema copy: Browse and select remote function modules from the flow node, explore the RFC catalog from a dedicated page, test destination connectivity, view interface schemas as JSON or XML, copy schemas in and out, and reuse imported objects as mapping schemas.
  • Unified Connections workspace: RFC, XI Proxy/PO and web-service connections are consolidated under a single tabbed Connections page in Operations, with inline BASIC credential creation for PO connections.
  • Web services (SOA) browser: Discover web services, view a WSDL outline, generate sample requests, and selectively import chosen services into MIP.
  • ActiveMQ connector: Sender and receiver nodes in the flow designer, so flows can produce and consume on ActiveMQ brokers.
  • Kafka topic monitoring and consumer tuning: A new Operations page shows topics and their message counts with fast cache-first loading; consumer start nodes gain a commit-strategy selector, a RECEIVED status breakdown, and an expanded, validated Advanced tuning tab.
  • Per-flow log retention: A new Flow Retention page in Database Management sets log-retention rules for individual flows, with clear hints and truncation warnings.
  • Unified credential management and service users: Credential management is streamlined into the edit pages with per-credential enable/disable, staged create-with-credentials, consumer OAuth2 binding, and single-mode hardening; service users get two role-split panels and a secure password generator with one-time reveal.
  • Resource pickers and safeguards in editors: Cascading Flow-to-Resource pickers to insert XSLT, XSD and flow scripts, a client-side XSD sample-XML generator, and save controls that warn when you edit resources used by live flows.
  • EDI converter and schema management: A unified EDI converter connector node plus an EDI Schemas management page.
  • OFTP2 connection usage modal: OFTP2 connections show a where-used modal listing the dependent flows and certificates/keystores before you change or delete them.
  • Flow designer productivity: Multi-node selection with Ctrl+click and group copy/paste, a multi-line SQL editor for JDBC query fields, and clearer Multicast guidance that warns (but no longer blocks) when a non-Direct connector follows a Multicast.
  • Clearer, more resilient monitoring: Loading spinners while payloads and full logs load, auto-focus on the failed connector in Show Full Log, refresh that recomputes the selected time range, an interface column and a MIP Message ID column for XI flow endpoints, and restored Sync Jobs and Component Health with a disk-usage column and per-node rows.
  • Richer activity history: The Activity panel shows human-readable node and field change details for flow updates, hiding empty fields and grouping cleared values.
  • "No Payload" log level on the Endpoints selector, and a REST Multipart Attachment Logging toggle.
  • OpenShift support: Ready-to-use OpenShift manifests and multi-variant container images.

Bug Fixes​

  • The Consumer list in API management no longer shows "Nothing Found" over a populated list.
  • The XSD/WSDL schema designer generates correct schemas — fresh previews, SOAP 1.1 and 1.2 output, full input validation, and confirmation before destructive edits.
  • Per-flow overrides of global flow configuration are correctly block-only, with an accurate in-use warning on delete or disable.

Health Check​

New Features​

  • External system-health API: External systems and monitoring tools can pull a single consolidated MIP system-health snapshot over a secured endpoint — overall severity, the status of each component and node, current resource usage and detected anomalies. Access works with either an API token or a username and password, and the response clearly lists any section that could not be read instead of silently omitting it.
  • Per-component health overview: A new overview reports the live state of every MIP component — healthy, degraded, recovering, unhealthy or unknown — with its response time and check interval, broken out for each backend instance, each search-cluster node and each database instance, flagging high disk usage as a warning before it becomes an outage.
  • Component version details: The system-information view now shows real version numbers for MIP components — backend server, web server and running container images — in containerized deployments where these previously showed placeholder values.
  • OpenShift deployment support: MIP Health Check can be deployed on OpenShift/ARO using native manifests and rootless image variants, with a ready-made monitoring and dashboard integration.
  • Fast-failing health probes: Dedicated readiness and liveness probes let an orchestrator detect an unhealthy or database-unreachable instance within seconds, and allow the service to start even while the database is temporarily down instead of getting stuck in a restart loop.

Bug Fixes​

  • System memory usage is reported accurately; it previously counted reclaimable buffer/cache as used and could read as high as ~96% on a largely idle machine.
  • Time-window filtering of health data honors an explicit timezone offset and a configurable business time zone, so charts line up with the intended local day.

This is a Release Candidate (pre-release); after validation the final 1.17.0 release will be published. Previous: 1.16.0