FME Flow

If you have a workspace app that requires authorisation, which runs a workspace that makes a http call to the FME Rest API, is it possible to use the logged in user to authenticate the http request?

I’ve seen a few ideas about how to deliver a data download URL after an automation as well as this ideaRun Workspace in an Automation using Data Download service | Community Has anyone used this sort of process before?Automation workspaces writes a file to a FME Resource location The last workspace in the automation triggers an external action which is actually a HTTP Request that triggers an FME Workspace with a filename and email parameter. Filename parameter is used as input for a filecopywriter and is published as a data download service, workspace gets the data download link and emails it to the email address supplied by the parameterAny possible pitfalls? Anything simpler that I’m missing?

I am using FME Flow 2026.1.1. The trigger for the automation I created is Resource of Network Directory (updated). I have a folder that FME Flow can read/write to/from. This is the Directory to Watch. This folder is in Resources.Events to Watch for is set to CREATE.The Poll Interval is 1 minutes. The automation is supposed to trigger when a file is added to the folder. But it is not triggering. Is there something else I need to do when using this type of trigger or maybe another trigger? I tried setting the Events to Watch for to CREATE, MODIFY, DELETE and the automation still did not trigger. Any ideas on how to get a Directory Watch to trigger when a file is added to a folder? 

I am working with FME Flow 2025.2.4I have an automation with several Workspace Actions. Parameters are passed between the workspaces using JSON automation parameters.One of the Workspace Actions needs to be dynamic, so I pass the workspace name through a string automation parameter and the remaining parameters through a JSON automation parameter.Everything works correctly when I use only standard Workspace Actions. However, as soon as I replace one of them with a Dynamic Workspace Action, the JSON can no longer be parsed inside the workspace. It appears that FME Flow automatically deserializes the JSON parameter before it reaches the workspace.As a result, the value is no longer a valid JSON string, and transformers such as JSONExtractor or JSONFlattener fail when trying to parse it.For example, a parameter that arrives as:JSON{"umgebung": "tst","parameters": {"port": 54322}}is transformed into something like:Plain Text{ umgebung : tst, parameters : { port: 54322 } } which is no longer

I recently come across some workspaces that I can't seem to upload anymore to FME Flow (2024.2). I already found the culprit, but still would like to understand why this happens and how to fix it permanently.Publishing a workspace gives an error. Dragging it into FME Flow will give a screen and that will get stuck and dragging multiple workspaces at once will upload all of them except the one not working and also not giving any feedback that one has not been uploaded. Eventually I found that pressing the Edit button behind the Job submitter gives more options and that there is something red. Easily to miss…So settings this to empty or another one gives me the ability to upload again. But next time it is again on the red one. Why does it keep remembering a FeatureReader that is not in the workspace again. I found also some documentation about what the use is of this feature, but I'm not sure I completely understand. If somebody could explain in a few sentences that would be great.

Hi PeopleI have a bit of a weird scenario that has stumped our IT team and our FME consultants.Our setup is that we have FME Flow Hosted as the controller which passes the jobs to an on-premise FME Flow server.  Whenever the on-premise server is restarted (e.g. Windows patching) Flow Hosted loses the connection.While that is to be expected, the connection doesn’t come back on its own when the on-premise engines are back on-line, nor can it be restarted (it hangs on initialising).  The only way to reconnect the Hosted service to the Remote Engines is to complete drop the Remote Engine configuration and recreate it.Has anyone else come across this or, more hopefully, fixed it?Thanks

Hi, I don’t believe it is from my look at FME Flow’s implementation of Tomcat, but can someone at Safe confirm that the critical Tomcat CVE-2026-86248 does not impact FME. Also, I assume this will mean an upgrade to Tomcat in the next FME Flow version?Thanks,Todd

As per this post last year, Salesforce is retiring the SOAP API login().  Any news from FME on when the Salesforce readers and writers will be updated?  

Hello, Does anyone know when the issue with MCP Flow authentication with Entra will be fixed? The apparent reason is that FME does not expose the scopes that Entra expects.  FME says the fix is to use an alternate SAML provider such as Auth0.  There is a work around, ish, sharing it here in case it can be helpful to others. Since switching SAML providers is not an option in my case I was about to give up and just wait for SAFE to fix this.  However, I wondered about simply connecting Auth0 to Entra and passing things along (federated).  ChatGPT seemed enthusiastic about that option so I decided to give it a go with it’s assistance.  My requirement is that users must be authenticated via Entra, adding AuthO as an IdP in FME Flow was not an option, no exceptions.  Sparing the details, here’s the basic setup. (I already had an Entra SAML setup working for Flow user login): Auth0 > Entra Enterprise connection for federation (free tier allows for 1).   Follow this tutorial for the Auth0

HelloWe have FME Server v2022.2.2 installed on Windows 2022 server and need to apply manual patch for security vulnerability, as per FME alert / articlehttps://support.safe.com/hc/en-us/articles/31265482270349-Security-Update-FME-Flow-Privilege-Escalation-Vulnerability#h_01JB2AENHZEBG9VCCFBJFTWX1RThe instructions include a note:*When moving/overwriting files, please ensure that the original permissions of the file being overwritten are preserved, this is ./opt/fmeflow.How best to do this?  We were planning on a simple copy and paste of the files themselves to the various directories, however this may overwrite existing permissions.  What method of copy & replace files is recommended, it would be helpful if the windows command or powershell script could be provided to assist Please adviseThanks 

Apparently since fixing this problem https://support.safe.com/hc/en-us/articles/25407612007949-Known-Issue-FME-Flow-Apache-Tomcat-Vulnerability-with-X-Frame-Options-headerIn FME Flow 2025.1, embedding FME Flow content to other pages is disabled entirely by default.  Is there any chance to allow embedding from known sources or is embedding considered so risky that this would not be even considered? 

We are working through the setup of a service user for a project, intended to only be able to create and run its own Data Virtualization APIs and nothing else.  The aim is to avoid excessive permissions. I have been searching for a list of recommended or minimum permissions but have not yet established that, so I’m putting the question out here.We have granted:Create/manage for data virtualization Run jobs Access Resources/DataWe’re looking into other permissions.We’ll work through it but I am definitely thinking that documentation of this would be handy from the Safe point of view to save time, if possible.  Possibly as an addition to https://support.safe.com/hc/en-us/articles/36098228739469-Secure-Data-Virtualization-Endpoints-with-Authentication 

Hi everyone, I am having the strangest issue with FME Flow. I have a Form that process some data and then creates a pdf file. It works just fine locally with a UNC path. My issue is Flow, when I publish it, i get a permission denied, but the strange thing is it has no issue writing .txt file to that folder and to make it even more strange, it worked correctly two times out of the 50 tries I did changing things. I have even tried adding this UNC path as a resource in Flow and same result. Has anyone had such issues before? Any suggestions? 

I was checking through the list of possible job status for an asynchronous (Data Virtualization) workflow I have going and didn’t find a definition for ‘complete’, or a comparison of that value to ‘success’.  I was referring to the REST API V4 documentation here: https://support.safe.com/hc/en-us/articles/38420159907213-Submitting-a-Job-through-the-FME-Flow-REST-API-V4 Would someone with familiarity around those statuses be able to clarify the meaning of ‘complete’ and how it contrasts to ‘success.  AI relented from guessing and admitted it didn’t know. An update to this is that ‘complete’ is not listed in the OpenAPI spec for the FME Flow REST API V4, so at this stage is the documentation incorrect?https://docs.safe.com/fme/html/fmeapiv4/docs/schema/static.json "status": { "type": "string", "description": "The current status of this job", "enum": [ "queued", "running", "success", "failure",

I have a workflow that reads an Excel file (attached image ExistingSourceDataset_XLSXR_image.png) then uses an AttributeManager transformer to modify the data, then writes the results to a SQL server database (attached image sampleFMWimage.png). Works great.I’ve been asked if it is possible to replace the hard coded Excel filename in the ExcelReader with a prompt for the filename. I thought it would be easy. I was wrong. My original thought was create a User Parameter (I’ll call it excelFilename) and then replace the existing source dataset with that User Parameter. Found no way to do that. Has anyone done something similar? Or know how to make that work?

About half a year go we set up a API Token thats being used by an external Web Application which triggers jobs on our FME Flow environment. This worked all fine till we upgraded to 2026.2 this summer.The right's we had set up were:Jobs AccessRecources Access (only to the required resources folder)Run Workspace AccessSince the upgrade to 2026.2 nothing was changed by us in these settings. The web application can still start jobs as needed. But any request of GET /jobs/{id} , /jobs/{id}/result en /jobs/{id}/log gives a 403 error now. The only way i found around it is to give the API Token the Jobs Manage permisson.But this feels wrong to me. If i read the documentation this would give the token access to all users jobs instead of only the ones started through this API Token. I can't find any mention about changes concerning this (only about the Tokens for FME Flow Apps which we are not using). Anybody has any idea what could be the cause here? Am i overlooking something?

Hi,Looking through updating our workflows to accomodate for the new v4 API Endpoints and getting the error message “Content type 'application/octet-stream' not supported” when testing out the new POST /repositories/{repository}/items/upload Is the content-type application/octet-stream no longer supported as part of these API updates and if so any suggestion for the alternate content-type to use when uploading .fmw files Thanks 

Every night we run make backup of the FME FLow (2025.2.2) system configuration. Since a couple of days this results in an error: | An error occurred: Topic SYSTEMEVENT_ERROR_LOG does not exist..No backup is created. When trying to start the backup manually, the result is the same Does anybody have an idea of what may cause this?

I have automations owned by ‘admin’ (superuser).If I come with another user that has full access, stop the automation make changes, save and enable again…. I find the next scheduled run under the Job logs, it is then run under the username that last edited it.this can lead to issues if the dependent components like run workspace > connections etc are also not granted permissions to that userI’m using 2026,1 … was this always the case or does the owner of the automation run the job? Maybe this change in the FME REST api v4 ? 

Hi,I noticed something unexpected today with deployment parameters that I would like to hear others opinion on.The situation at hand is that we have several automations running on FME Flow. In these automations an external action is present to send an email upon success. For the 'Email_To’ parameter of this email action, a deployment parameter is being used.It thus currently looks something like this;And then in an automation it would look like this;Then I had this automation running, and if it gets triggered, the email is send like intended.But now I wanted to remove one recipient of the emails. Therefore, I changed the value of the deployment parameter email_to from 'email_adress_1,email_address_2’ to 'email_adress_1’;Now for the unexpected behaviour. I noticed that after changing the value of the email_to deployment parameter, if I would again manually trigger the (still running) automation, I noticed the email would still be send to both the email adresses, where I expected it to h

Hi there,If we change the URL of our FME instance all the redirect URIs of our OAuth web services will be pointing to the old URL.Is there a way to bulk update them, either through a setting or an API call? Thanks,Marc

I have an automation that runs every 15 minutes on a single engine with a ‘repeat on interval’ schedule trigger. Usually, it takes less than 10 minutes but sometimes it can take more than 15 minutes and once a week on mode overwrite it takes about 25 minutes. If one of the workspaces of the automation is still in progress, is there a way to skip the trigger for the next automation?

I want to create custom UI for a flow app like create button. I want to build some sort of app that gave user more creative UI and good experience whether it is by HTML or CSS.

Hi Folks,Almost weekend and we got a bit of a strange situation. We have a Flow App that is being used by most of our users. They upload a GeoJSON and select if they want Aerial Imagery or Topographic Background. The result is a generated PDF Map.This process works fine for all but one user. When this user uploads a GeoJSON we see in the log the following error:90 2026-8-7 15:44:02 | GEOJSON Reader: The file 'C:\ProgramData\Safe Software\FMEFlow\resources\system\temp\upload\[Projectname]\[Workspacename]\auth-user\[USERNAME]\508a819f-ee99-bb88-9785-82b3a169082e\[GeoJSONname]' does not exist91 2026-8-7 15:44:02 | An error has occurred. Check the logfile above for detailsIf any other user uses the same GeoJSON file the process works fine.We checked the following:User rights are identical The GeoJSON is valid (others can use it fine) We checked the timestamps in the DataUpload log with the Job Log and the file is uploaded before it's read. We checked with the user where it doesnt work if t

Hello!We are using Flow FME 2026.1.2. I would like an overview of when most automations trigger. Unfortunately, there isn't much information about it in the documentation, so I did some research and found this post:I tried it and now I'm experiencing some really strange behaviour. Until today, I could run all the API requests I ever needed and get results. I can even run FMEURL/fmeapiv4/schedules?limit=100&offset. What I can’t run is the solution in the thread:FMEURL/fmeapiv4/schedules?limit=100&offset=0&sourceType=automationAs soon as I add 'sourceType' as a parameter, I get the following result:{"message":"Unauthorized request by user USERNAME due to lack of proper permissions or the object does not exist."} Do I need special privileges to use this parameter? Or is it just bugged? Is there a workaround to achieve my goal of creating an overview of automation schedules? Many thanks in advance for your help.Kind regards, Henning