tl;dr: The rules vary, check your intended journal's policies
I'm going to interpret your question as asking about what's acceptable in publishing, what issues you have to consider when republishing work and what problems you might run into.
You have three considerations here (self-plagiarism (1), duplicate publication (2) and copyright (3)); the boundaries between these can become somewhat blurred. Typical conventions vary from field to field and specifics will vary from journal to journal and publisher to publisher. Let's look at each consideration in turn, from a practical perspective the section on journal rules is probably most important to you. This is because, although it's easy to obviate self-plagiarism with adequate referencing (see section 1), this doesn't guarantee compliance with journal duplicate publication policies (section 2).
1) Publishing/scholarly ethics
(For ethical background on duplicate publication, see: 1, 2, 3.)
You should always cite original sources, and be specific where you've lifted quotes/figures from them; just as if you were taking from someone else's work. This is enough to obviate the risk of valid accusations of self plagiarism, but you still need to consider issues like duplicate publication (see links above and the section 2) and copyright (section 3). If work has been made publicly available (either by you, or by someone else), you should consider (for the purposes of referencing) that it has been published, however (in)formal or (non-) academic the context.
The only exceptions I can think of for this are: preprints, where the preprint server will be updated with a link to the version of record instead; publishing work that was included in a PhD thesis (strikes me as odd that a reference/note isn't required in these cases, but typically one isn't); publishing work that was presented at a conference without publishing a proceedings paper (even if the presentation abstract is publicly available). Link 1 goes into a bit more detail on duplicate publication.
In your case, you should cite/reference the online posts. How/when you do this depends on exactly what form the work takes, but a line in the introduction along the lines of the following may well be suitable:
This piece grew out of some posts the author(s) made on [forum][reference post(s)]; here, the author(s) have revised and expanded these posts into a reflective piece on the topic of...
In the case of any doubts about the acceptability of such a submission, mention it in your cover letter. This is probably a good CYA policy, even if you're sure what you're doing is allowed.
2) Journal-specific rules
Every journal will have its own specific rules on prior publication, blending the considerations under publishing ethics and copyright (discussed below) and considerations such as field-specific practices and editorial considerations. These dictate what is or isn't acceptably in a particular journal. You will have to look at your intended journal's policies yourself to find out what they are, if you're not sure how you fall under them, their editorial office will be able to guide you.
In general, putting a thesis online is almost always OK and use of preprint servers is becoming increasingly widespread. Other online publication, is not necessarily allowed. Let's look at some specific examples (emphasis added):
The Nature Research journals are happy to consider submissions containing material that has previously formed, and continues to form, part of an online scientific collaboration such as a wiki or blog, provided that the information has not been publicised outside the scientific community, and is not publicised until the publication date of the work in a Nature Research journal.
In this case, it would seem to depend on whether or not the forum could be considered an academic collaboration or not
The Science Journals will not consider any original research paper or component of a research paper that has been published or is under consideration for publication elsewhere. Distribution on the Internet may be considered prior publication and may compromise the originality of the paper as a submission to a Science Journal.
ACS ethical guidelines (note, I don't think these are their publishing policies, but good practice guidelines issued as a professional organisation)
Scientists should, however, be aware that disclosure of research results in the public press or in an electronic database or bulletin board might be considered by a journal editor as equivalent to a preliminary communication in the scientific literature
From Wiley's best practice guidelines
Journals from different disciplines vary in their approach to pre-print servers. Many biomedical journals would consider posting an article to a pre-print server to render any subsequent journal publication redundant. Thus an
article submitted for consideration after having been posted to a pre-print server would be rejected. However, many researchers working in physics, mathematics, computer science, quantitative biology, quantitative finance and
statistics post their articles to arXiv before submitting an article successfully to a journal for peer review and publication. Journals should establish a policy about pre-print servers and declare this in their instructions for authors. Any previous publication should be disclosed in the paper.
The following types of “prior publication” do not present cause for concerns about duplicate or redundant publication:
Results in databases and clinical trials registries (data
without interpretation, discussion, context or conclusions
in the form of tables and text to describe data/
Dissertations and theses in university archives.
I won't give any more examples, because without knowing what field you work in for someone to give specific advice (mention of a "reflective piece" suggests it not STEM?), I don't think it would add much to the above.
If you aren't sure whether your use falls under the rules for your journal, you can contact the editorial office. You should also mention in the cover letter when you submit, eg. a sentence paragraph along the lines of
This work is an extension on materials previously posted to [forum] and, consequently, makes heavy reuse of this material. We believe this is acceptable under [journal]'s rules on prior publication.
You will need to confirm that you have the rights to reuse your posts. Double check the small print of the forum in question to make sure you didn't transfer copyright when you posted them (hopefully unlikely!). If you did, you will need to obtain permission to reuse them from the from the forum operator and provide this to the journal during submission.