FME Flow

After upgrading from FME Flow 2024.2.3 to 2025.2.4, the labels (titles) under Gallery Apps are now rendered as blue, underlined hyperlinks.Previously, the labels were displayed as plain black text without underline, even though the entire app tile was clickable.2025.2.4:2024.2.3:Are there any built-in configuration options or recommended approaches to control the styling of Gallery App labels? We are aware that other organizations have experienced the same issue after upgrading from the same versions.

Hey everyone! I wanted to share something I’ve been working on for a while.When I’m debugging our own FME Flow, or one from a supported user, I often need to dig into the log files to track down errors or evidence of what went wrong. This can be really time-consuming: there are many files to open, and some of them are very large. Most of the time I end up just jumping between files and scrolling to the right timestamp, instead of actually focusing on what happened. To make this easier, I built an FME workspace called FMEFlowLogInvestigator.This workspace can be published on FME Flow with the Data Streaming Service. It takes four user parameters:

If an automation is triggered by a schedule, you can obtain the schedule properties via the Schedules list.  https://docs.safe.com/fme/html/FME_REST/apidoc/v3/index.html#!/schedules/list If an automation is triggered by a topic, how do you obtain the name(s) of the topic(s) that trigger that automation?If an automation posts to a topic, how do you obtain the name(s) of the topic(s) that automation posts to?

I’ve noticed a few issues since upgrading FME Form & Flow (FME Flow 2026.1 Build 26103 - win64) that seem to show up when I’m also developing in FME Form while remoted into our FME test VM.This is how I usually work, mainly because some jobs take a while to run and if I do them locally my PC can go to sleep after 5 minutes and the jobs fail. This setup has never had an issue in the past. Since the upgrade though, running Form this way can cause problems with FME Flow on the same VM. When I’m running heavier workbenches and CPU hits 100%, Flow can become unresponsive or look like it’s soft‑locked. This never happened before the upgrade.One thing I’ve noticed is that this seems to mostly affect SSO sessions. My theory is that when CPU is maxed out, Flow struggles to grab required info for SSO before it expires. Just a guess.

It seems the issue described in the related topic, where the job variables show in the result, instead of the values, has never been resolved. Or the issue has returned in version 2025.2.4.As a workaround I can send an email to a Teams-channel. But I looks a lot better when I send an adaptive card to the channel, using a Workflow in that channel.But as described in the related topic, when I send a adaptive card with a HTTP Request in an Automation, a job variable like {automation.name} just shows as {automation.name} in the result. Of course I would like it to show the name of the automation.Is there a way to make this work? Or is this a bug and I just have to wait when it will be fixed?

Can the dynamic parameters in FME Flow Apps, as described here:Dynamic Parameter Configurations in FME Flow Apps – FME Support Centerbe made to cascade, i.e. so the value chosen in one parameter affects the choices available in another?How does one structure the JSON to do this?

Hi all, we are on Flow 2025.2 and we use Geometry Parameter for quite a few ws apps. Now we are living with the bug FMEFLOW-26285Changes to the default basemap of a geometry user parameter under the Parameter Defaults when creating/editing a Workspace app do not get applied to the Workspace app. Users accessing the workspace app will see that the default basemap is unchanged. Any tips for a workaround? Can one change the default using javascript in the webpage possibly?

Hi, I am a software developer, but don’t know much about FME yet. Please forgive my ignorance :-)I was asked to create a simple wrapper around the REST API v4 that lets a user start a job, and download the results as a zip file.Generating a token, submitting a job, and getting the job’s status is fairly straightforward, but I am at a loss when it comes to downloading the results. I see that there is another API /fmedatadownload, but it seems that the token I use when submitting the job does not work here. I tried passing it in the Authorization header, as a query parameter, and I also tried Basic authentication, but all of these return HTTP 401I am sure this must be very simple, but I am probably overlooking the obvious here.

Hey all, After updating my FME Form to version 2025.2.2 & updating the ArcGISConnector package one of my FME Flows has stopped operating in the same way. I have an ArcGISOnlineConnector Transformer which updates a hosted CSV item on an ESRI Enterprise system. This had been previously working well & updating each row within the CSV and the CSV on Enterprise showing as updated after the flow had completed. Following the update the Flow now only updates the First & Last feature according to the log. The CSV on the Enterprise does also not show as having been updated when the flow has completed. This worfklow does still work on FME Form. Previously FME Flow went through each feature on by one like the second setFME Flow:125 2026-4-17 08:52:50 | ArcGISOnlineConnector_2_CALLER (PythonFactory): Processed 1 of 16,927 features126 2026-4-17 08:52:51 | ArcGISOnlineConnector_2_CALLER (PythonFactory): Processe

Hi all,I have a fault tolerant deployment of FME Flow 2026.1 in our private AWS account. Our security team requires us to limit the number of ports to be opened between WebUI/Core and Engine hosts (eg. least privilege).Is there a best practice rule/formula in calculating the number of ports required to be opened depending on the number of engines per host and the number of hosts in your deployment and the starting port number? ThanksExample:3 engines / host and 2 hosts = 6 enginesIs there a redundant factor that needs to be applied? eg. 2 timesTherefore 3 engines x 2 hosts x 2 redundant factor = 12 TCP ports are required.The starting port number should be from 50000 fmeFlowConfig.txtFME_SERVER_PORT_POOL=50000-50011

I have several automated jobs that use either a startup or shutdown python script.  These scripts run fine in workbench and even in the ESRI Python interpreter on the Flow machine.  The issue arises when I run them from Flow.  When it gets to the script it will output expected products until the exportToPDF command at which point it fails.  No errors are indicated anywhere except in the Engine Log where this is displayed: WARN     localhost_Engine2   393562 : Process "localhost_Engine2" ended unexpectedly. Being restarted on attempt 1…Example of script:import arcpy, os, sys, time, shutilfrom datetime import datetimesdpFolder = r"****OBSCURED****"exportFolder = os.path.join(sdpFolder, "SiteDevPlanFiles")date = time.strftime("%Y-%m-%d")dateFolder = os.path.join(exportFolder, date) archiveFolder = os.path.join(sdpFolder, "_Archive")zipExportFolder = exp

Most REST API’s I have worked with work in a similar way with regards to authorization: Based on some credentials, you generate a token, and use that token for subsequent requests. You can usually create as many tokens as you want using the same request. What I noticed in FME (but please correct me if I’m wrong), is that a token must have a unique name, and when using the same request, you get a response that says Specified user token already exists. Of course I can update an existing token, or delete it and create a new one, but that means that in most cases, I need two requests instead of one.Another thing I noticed is that expired tokens are not deleted automatically: A request to /fmeapiv4/tokens showed me a token that apparently I generated 7 years ago! I am curious about the reasoning behind this way of using tokens. I mean, why does a token have to have a name, and why would you keep generated tokens forever?

Has anyone else run into issues with how things are displayed in FME Flow 2025.2.1 Examples: Engine Assignment Rules - we have two hosts which differ in name by suffix only. The way engines are displayed on the Engine Assignment Rules tab means you can’t quickly tell which engines are assigned as the critical bit is displayed as … (the full names are available on mouseover) The hostnames displayed in No proxy also have their urls partially replaced with …. so there appears to be no way to actually check via the user interface what values are entered here. In 2024.1 these values were displayed in full 

 Using $ref in FME Flow Data Virtualization APIs for Deduplication This might be an imprecise post, but my question is whether FME Data Virtualization follows the exported API specification or the visual editor’s interpretation - and why. I need guidance on restructuring my specifications: Data Virtualization seems to accept arrays of schema references but silently drops them, replacing them with blank references in properties but losing the json-type array in endpoint responses.This inconsistent handling of $ref makes it unclear how Data Virtualization processes schema references — especially for arrays, since the visual editor only supports $ref for objects. Thanks for taking the time in advance!EDIT: Im am currently using FME 2025.2.4 Build 25835 The Use Case of Arrays with $refHere is an example usage of Arrays of Referenced Objects from the OpenAPI s

Hello everyone,We are encountering an issue with FME Flow 2025.2.1 where jobs stop unexpectedly without any error message.Scope / Versions✅ Issue occurs in: FME Flow 2025.2.1 ❌ Issue does NOT occur in: FME Flow 2024.1.2.1 FME Form 2025.2.1 This suggests the problem is specific to FME Flow 2025.2.1 and not related to the workspace or FME Form itself.Observed behaviorThe job stops abruptly without any error reported in the FME Flow UI or job log. Enabling Debug logging in FME does not provide any additional insights. The same workspace runs successfully in: FME Form 2025.2.1 FME Flow 2024.1.2.1 FME Flow loggingWe reviewed additional server-side logs.In fmeprocessmonitorengine.log, we consistently see that the engine en

I started protecting my workspaces with passwords because they sometimes contain sensitive informations (like embedded connections) and I noticed a weird issue when I uploaded them on FME Flow 2024.2.5.I use the parameter FME_MF_NAME in different places, for naming different outputs. As soon I added the password, the workspaces started to fail. Notably I have one workflow which requires launching multiple workspaces sequentially, the main one launching the others with FMEFlowJobSubmitter. All the protected workspaces have their FME_MF_NAME overwrited by the FME Flow. All the logs show the same message upstream, something like this:FME Configuration: FME_MF_NAME is 'wb-tmpXXXX'Which is weird because the parameter of the workspace is correctly set with a message like this:`--FME_MF_NAME' `myworkspacename.fmw'Using the parameter FME_MF_NAME during the job gives 'wb-tmpXXXX' instead of ‘myworkspacename.fmw'.I can find workar

Hello,I believe I’ve encountered a bug in FME Flow 2023 and would like to know if others have experienced something similar.I recently updated a workspace by changing the recipients in an email transformer and overwriting the file on FME Flow. This workspace is triggered by two separate schedules: one for Mondays and one for Thursdays. Both schedules are identical except for the day they run.Surprisingly, when the Thursday schedule ran, the email was sent to the old recipients (prior to the update). However, when the Monday schedule ran, it correctly used the updated recipients.I have verified the following:The logs confirm that both schedules are calling the exact same workspace file in C:\FMEFlowShare\repositories. I manually checked the workspace on the server, and the parameters are definitely correct.Why would one schedule seemingly use a cached or "ghost" version of the

Hello community,for a few days, I have been struggling with our FME Flow. It occasionally throwsand I am the admin, I asked myself and… crickets.The only change we did recently is a Python library install.FME Flow gets in this error communicating state and then recovers. We tried restarting FME Flow, it likely made some difference (I am not sure whether it indeed fixed the problem even for a short period of time as FME Flow works between these messages anyway), but it didn’t last.This is FME Flow 2024.0 build 24612.Any ideas will be much appreciated.Thank you :)

In Flow 2025.1 the message user parameter type would show with an info icon and blue background in a workspace app.  In 2026.1 there is no colour behind the message, just the text and the plain white background.  I cannot show much here but this is what I’m seeing - the top line should have the blue background and info icon:Interestingly in the workspace app editor the blue background does show for messages there. 

Dear FME users,Already for days we're trying to get an automation to work with a HTTPCaller using certificates.The HTTPcaller works when we put in the HTTPCaller the option 'Verify HTTPS Certificates' to 'No’, see printscreen below.However when we publish the workbench to FME Flow and run it from an automation we are getting the following two errors:Aanroep_gateway_lijst_ftp (HTTPFactory): [1]: HTTP transfer error: Problem with the local SSL certificateAanroep_gateway_lijst_ftp (HTTPFactory): An error occurred while accessing the URL 'https://tst….

Hi,I need to make the Esri Arcgis connector available for all users within a APP-V packages, however these connectors install under the app data folder. Is there any way to install these connectors on a program files or data folder so that all the users can make usage of the Hub connectors. Is there a way to change this path? Thanks in advance. I already tried to copy the files to the program data however FME doesn't see the files and is unable to access the connectors. Yours sincerely.  

For one of the first versions of Data Virtualization, it is required to create a work-around for IIS. Unfortunately it is not possible to react on the knowledge article directly anymore.I just updated to build 25835 and after a quick test I see the work-around is still required. From which version onwards is this not necessary any more?

Dear All starting using FME Flow - MCP Server,I have an existing FME workspace performing semantic search using Microsoft Copilot Retrieval API searching against SharePoint Document Library, which was used with a FME Flow App (Semantic Document Search Engine). It will return HTML based search results with generative AI vibe-coded UI. It has been working quite well for our end users.Now, since we are all very excited about potentials of FME Flow - MCP Server. I have made a copy of my workspace and convert the out to JSON writer or TEXT writer to streaming JSON of the semantic search results.However, Claude AI seems to only be able to get returns of job submission log from the MCP tool: { "id": 182, "featureOutputCount": 2, "requesterHost": "X.X.X.X", "requesterResultPort": 57366, "resultDatasetDownloadUrl": null, "status": "SUCCESS", "statusMessage": "Translation Successful",

Hi,I am using FME 2026.1, trying to create a map using AttributeCreator+ PDFStyler.However, when I set my label height to for example, Arial,5, my text comes out bold. When I select Arial,5,BOLD, my text comes back looking regular. Is that a bug just for me? I was stuck on it for ages trying to figure it out. Yes, it is also the case when i set a font size like 5 (not from an attribute).I am wondering if anyone else has this issue.   

Testing after upgrading preprod to FME 2025.2.1 and it appears that an issue that would previously (FME Flow 2024.1) be reported back from the database, i.e. incorrect user/incorrect syntax do not appear in the log filee.g. running a workspace with an SQLExecutor with a query with incorrect syntax in FME Form 2025.2.1 results in an error, e.g.“java.sql.SQLException: Line 1: Incorrect syntax near '}'.”No such message appears when running the workspace in FME Flow. The workspace fails but no SQLException is returnedThis appears to be the same whether you connect with using JDBC or Microsoft SQL Server (JDBC).