EXCEPTION: stale element reference: element is not attached to the page document - selenium

Question

My HTML-Page contains (among other stuff) this bit:
<ol id="links">
<li id="links_1">stackoverflow</li>
</ol>
In my code to test the page, I then do:
el←FindElementById'links_1'
(ACTIONS.MoveToElement el).Build.Perform
and this crashes with
EXCEPTION: stale element reference: element is not attached to the page document
(Session info: chrome=81.0.4044.129)
However, this error doesn't seem to be justified, as the element is still alive:
el.Displayed
1
el.Text
stackoverflow
el.Location
{X=56,Y=282}
How can I fix this problem?
(The environment is APL. I've left out a few APL-details here because I feared they might be avoidable "distraction" from the core-issue)
During my research before posting, I saw the question stale element reference: element is not attached to the page document but it doesn't seem to apply:
as shown, I'm doing the FindElementById and access it immediately after finding it. The DOM doesn't change, page is static.
explanations about it no longer being part of the DOM do not apply: it is found and, as I attempted to show, I can access properties such as Displayed or Text.
I also went through the reference, but this explanation did not help.
Also, there is no looping going on and nothing changes the page. It's really straightforward: GoToUrl * Find * MoveToElement.
I removed the chromedriver-tag, as I can repro this with Firefox and geckodriver. However, with geckodriver, I get "EXCEPTION: Web element reference not seen before:" when I do MoveToElement - but I can do el.Click and access its properties in exactly the same way that as with Chrome.
Update: a coworker investigated this a bit deeper (beyond my comfort zone) and found that before throwing this "stale exception", there is (I'm not sure how to word this properly and where exactly it occurred) a 404-exception. I just know 404 as an HTTP Status-code - that's all it means to me. But clearly the browser was not asked to navigate anywhere, so I can't be related to an HTTP404. Does that perhaps ring some bells with anyone more familiar with the internals of WebDriver?

Answer 1

This was a complex and multi-layered problem, but when I finally removed all layers - it worked!
The key-factor which caused this problem:
due to a fault in the way I had written the test caused it to be executed twice - and the 2nd run always exposed the problem.
I had instantiated ACTIONS in the test-framework and was not aware of a critical feature: it continuously builds a chain of actions, any Build.Perform..-steps just added to that. Solution: create separate instances of every run (possibly, since it's fairly lightweight) or call ACTIONS.Reset(requires WebDriver4). I've never had issues with WD4 (although it still is in alpha) - but this gave the ultimate reason to switch!