Remember that Denmark only has 6 million inhabitants, and they didn't make the final. The number isn't that low all things considered.
Plus no idea if Australia's numbers are only considering the live broadcast of the show, or also the reprise in prime time, but these numbers aren't that low when you consider that Australia's most watched shows only get around 1 million in ratings, and that SBS's usual ratings specifically are below that number as well.
These ratings systems have historically lagged in counting online videos/video streaming and depending on the nation those in the sample may or may not count if they watched Eurovision by an online source rather than the terrestrial broadcast. For example in Australia it does not currently count if you watch a YouTube video or Facebook video, even if that video is of the same content also available through a different broadcast which is counted.
Also, those TV rating systems are not entirely accurate. They use a small sample of the broader population who have devices which log what they watch...for example, in the UK the ratings are determined by
12,000 people in 5,100 homes (while the total UK population is 67 million people).
The OzTam ratings in Australia has
1475 people determine what the 5 million Melburnians have watched and another 1475 people determine the same for the 5.3 million in Sydney, with both cities being widely considered among the top ten most diverse cities in the world.
These ratings systems are relatively accurate if something is of broad general appeal, but when this is not the case the data is increasingly inaccurate. In some nations Eurovision has a broad appeal, but in others it has heightened appeal for more specific audiences, such as those who have a strong link to Europe or with an interest in global musical culture, or LGBT+ culture (especially in nations where not being heterosexual is less socially and/or legally acceptable).