1. Control plane problems - these will manifest as latency or errors when calling the IAM service itself to make updates/additions/deletions of users/roles/policies/groups/etc. This is the most likely scenario if the dashboard names "IAM" specifically as the issue.
2. Propagation problems - problems with propagation will manifest as delays in seeing your control plane (IAM) actions reflected in the dataplane. For example, if you remove a user, but the user is still able to authenticate to AWS services for a lengthy period of time.
3. Dataplane problems - this will be problems with authentication or authorization to any and all AWS services. A widespread problem with authentication is less likely, but extremely bad, and will probably not be categorized in the dashboard as a problem only with IAM, since "IAM" is technically the name of the control plane.
My understanding is that authz/authn flows have not been affected. If authz/authn flows (which occur on every API call) were affected, I suspect the effect would be far more noticeable.
The biggest thing is what do you want to learn and how complicate do you need things to be. These two reInvent videos are really good:
They're a little old and might not talk about the newest and greatest things (like IAM Access Analyzer), but the basics of IAM are there and always stay the same.
If you give a better understanding of what you struggle with (basics, conditions, etc.), we could probably give better answers.
Here are some other reInvent videos I skimmed that look pretty good:
If you want something more hand-holding, A Cloud Guru courses are pretty good.
People rarely think about the true purpose and power of IAM and its equivalent services in other platforms. Its real function is to decouple teams from each other, which is what enables the platform to grow. IAM is the glue that federates services together. What developers see is a powerful, relatively low level RBAC/PBAC API for that, but through that API you can get a glimpse of just how agnostic IAM is to the information that it's managing - and how central it is to the rest of AWS. (Fun fact - you can actually use IAM to evaluate completely made up policies/principals/resources, because it's agnostic to what it's evaluating and the PDP API is available to everyone.)
Comparing this to Google or Azure, AWS IAM is architecturally superior - at least in terms of extensibility.