Your example is a deontic command, an instruction that something be done. It is equivalent to using the imperative mode directly:
Stop the process to investigate the issue.
For modals, English has:
- Exactly nine verbs that get full-time use: can, could, shall, should, may, might, will, would, must.
- A couple verbs that get part-time use as uninflectable modals, but which can also be used as finite verbs: need and dare.
- Various periphrastic verbs that are slowly graduating into being true modals, most notably including ought (to) and have to.
The central point that every native speaker intuitively knows (no matter whether they know that they know it) but that every non-native speaker must learn through study and practice, is that every single English modal verb has two distinct modes of use:
- A deontic mode that describe how the is supposed to be and that it needs changing.
- An epistemic mode that neutrally describes something known with varying likelihood but without making a value judgment.
When we are talking about an instruction manual, it is likely to follow (albeit usually unconsciously) the specialized senses of these verbs as spelled out in RFC 2119, which is the standards document that defines exact meanings for these too-easily-confused words. Just because a document fails to say that:
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL
NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and
"OPTIONAL" in this document are to be interpreted as described in
RFC 2119.
One should not assume that the definitions of those words defined in that document are not relevant to what you are reading, since that document merely formally codifies established practice common to all English speakers writing technical documents. It has been codified that there be no room for doubt in formal technical specifications.
The most important word is listed first in that RFC:
⒈ MUST This word, or the terms "REQUIRED" or "SHALL", mean that the
definition is an absolute requirement of the specification.
In other words, shall is exactly the same as must here, and it has the force of a command, an imperative. This is the deontic sense of those verbs, since it is saying how the world ought to be, not merely how it is going to be.
There is a famous opposition pair of examples that although of debated “correctness” (it seems that the distinction has never been one practised by the majority of native speakers) that illustrates the difference between these two linguistic modes.
The key is that once upon a time, some folks decided that shall is epistemic in the first person but deontic in the second and third, while will is deontic in the first person but epistemic in the second and third.
- I shall[EPISTEMIC] drown, no one will[EPISTEMIC] save me.
- I will[DEONTIC] drown, no one shall[DEONTIC] save me.
Under that interpretation, the first example, which is purely epistemic, might be the cry of a man fallen accidentally overboard and fearing for his life. He is describing likelihood of outcome. It is simply stating what is likely to happen. It is neutral of judgment in its forecast of the future.
In contrast, the second example uses (or at least, is purported to use) modals in a purely deontic sense: it is a suicide’s letter of intent. It is making a strong judgment on the world to come, insisting that the world be changed to match that intended outcome. Volition is strongly involved.
Recent research has concluded that this sort of distinction, in which the modality is determined according to whether the verb is in the first person or in other one, is one that most speakers do not follow, or at least, they do not do so naturally. It is a distinction that if it has ever existed at all in a descriptive rather than a prescriptive way is at most limited to a small set of native speakers from a particular region of the world and from a particular educational background.
For most of the world, will is the epistemic modal in all persons, while shall is the deontic modal in all persons.
This does not conflict with RFC 2119, since it specifically defines SHALL to be equivalent to MUST, and tells you that this is an absolute requirement of the specification. It is deontic: it is how the world is supposed to be, not just how it’s probably going to end up being.
See also the related questions of: