If the issue occurs sporadically, for layers you own, to new maps, and the layer otherwise works (i.e. not a recently uploaded csv that may have data conversion issues) you may be seeing service disruption. Easiest way is to add the layer to a new map and reload the page. See if it works sometimes but not always.
These approaches help rules out a number of conventional reasons a layer won't load in a map e.g. data issues, or the layer has become unshared.
With this, send an email to esri support identifying that it is a suspected service disruption issue in the subject title. Include the service URLs and ideally share publicly accessible ones so the initial triaging analyst can test. Let them know what you have tested i.e. help them rule out that it is not a conventional issue. Advanced captures like fiddler/http traffic or dev tools console errors may help. The service URL is very important even if it's not shared as it includes the server hive number.
This will then get passed to the Esri outage team for review and resolution. If it impacts a majority of users it will then be updated on the ArcGIS Online Health Dashboard. Not all service issues are placed here.
This is the fastest and only way end-users can notify and resolve issues like this. Esri will have automatic health checks in place that will pick up issues. You reaching out to support helps expedite this identification, track the users impacted, as well as sort out issues that are actually client configuration. The above information will help expedite it through the Esri triaging system. There are a lot of reasons a layer cannot be added to a map, and Esri support will want to check these for isolated reports before just shipping all issues to the outage team.
Hope this helps. I think posting here is less valuable as the community (esri staff and public) aren't in a great position to help outside of maybe looking at one of the conventional causes. The majority of users won't share the same server hives and by the time we're having active discussion here it'll likely be up on the health dashboard or even just resolved....
Feel free to share more on the configuration and behaviour you are seeing if you want confirmation that it's not a configuration issue.
