There is a report in Github github.com/indilib/indi/issues/1649 and here too indilib.org/forum/focusers-filter-wheels...-intermittently.html but as far as I can see there is no actual troubleshooting to replicate that.

if more users reports a problem and also with different setup, it is logic that there is a problem

You are totally right, I do not disagree with that. Even if only 1 person has the issue, is still an issue. For example I was the only one that reported issue with the iOptron's wifi mount drivers - it is even possible that I am the only one who use them - tried and replicate the issue, so Jasem tracked down the error and resolved it.

So I just quote Jasem github.com/indilib/indi/issues/1649#issuecomment-1084604284 "That's not exactly a step-by-step procedure to reproduce the issue. What equipment exactly did you have in your profile?". I am pretty sure that a lot of people would be more than happy to help out but if we can't replicate the issue, our hands are tight.

It seems that the ony way is to make shoots and check or analize the code that get the filter info and write it from a previous working version and the new one. I do this in my production of PHP/HTML code and i test it in much of all possibile scenarios before release a new version. obviously that can be some problems, but some

Hi
I also ran into the issue of no filter information in my subs last night - doing R,G and Ha subs. Pixinsights or DeepSkyStacker treats all the images as luminance.
I am at Kstars 3.6.0. Stellarmate 1.7.2

This Is open source anyways... anyone of us with enough skill can look at the code and fix It. Just to say...
Back on topic: I have just terminated 5 nights of captures changing broadband and narrowband filters during every night on two differenti rigs, and I didn't get a single filter tag missing sub. It's clear that the bug appears in very specifc circumnstances, otherwise we'd all experiment It: precise information and debug logs are essential for the developers to track down the issue and fix It.
I didn't experiment the bug, otherwise I woud have done It myself.

We would definitely fix if there was a way to reproduce this behavior consistently.

It has been a LONG time since I have done coding so looking at the open source code is probably not going to be useful

What I can do, now that I am aware of the intermittent issue, is to constantly keep track of my equipment and what I was doing at the time in order to try to reproduce the problem.
For example, the issue occurred the evening of Aug 10 beginning about 10PM when taking 30 180 second subs each for G,B & Ha imaging NGC 6995. The 6th blue sub had the FILTER field but no other subs did. All the images were fine and the system correctly did a meridian flip around 1:15AM CDT.

My setup was:
Ceestron EDGE 9.25 with 0.7x (primary) with ASI 1600mmPro camera
ZWO guidescope
EFW 8 position filter wheel connected thru the USB port on the camera with LRGBHaSiiOiiiK filters installed
GPSD dongle
Pegasus Focuscube
EQMOD mount
Astrometry
Openweather
Watchdog
running on a Raspberry PI 4 with latest Stellarmate software 1.7.2 connected to a Windows 10 system running Kstars 3.6.0 with Ethernet

Again, I will track when/if this happens again if I can isolate if particular components are involved.
Tahnks for the support.

Here the result from multiple nights using the Scheduler on the Helix Nebula.
As you can see on 193 frames, only 5 have the filter tag into the FITS header.
The setup was always the same, QHYCCD 183M + CFW3 and KStars 3.6.0 taken from the repository.

I just pushed some changes to INDI::CCD to enable more logging. So any drivers compiled from GIT now should have this logging information and if Filter Slot is unset for whatever reason, we'll at least have it logged so we can track down the issue. Not exactly a solution, but a step in the right direction.

Thanks. I'll try is in a next week, however I don't know what INDI::CCD means.

It just means you need to compile from GIT. So

update_indi_core

then

update_indi_drivers