Most journal articles aren't that important. Reading them (or not reading them) has no real impact on your research.
But, occasionally, you will find a paper that just fits. It's relevant. It's high quality. And most importantly, it influences the way you think about or go about your own research.
There will be relatively few of these (perhaps 5 or 10 in total during your whole PhD) but the impact they have is massive.
Once you've found one of these papers, (let's call it paper A) you can use it as a seed to grow the rest of your bibliography. Here's how...
First, look at the references in paper A. If the work is highly relevant to your own, the chances are they will cite other sources that are relevant to you. They have already read through the relevant literature and are telling you where to look.
Then, look up the authors of paper A; what else have they published? What do their departmental or LinkedIn profiles say they are working on now? This is not only to find other potentially relevant articles, but also to get to know who is working on similar topics (the field consists of people, and you need to know who they are).
Finally, look at who else has cited paper A. Most academic search engines and many journal pages include "cited by" information for each source. Anyone else doing similar work to you is likely to cite many of the same sources, so this is a good way to find results that you may not have found through a search engine.
This final step can also give you an indication of how paper A has influenced the field. It's also a good step to repeat to find out if anything relevant has been published recently (after you submit your thesis but before your defence, for example)
If you repeat this process using the best papers you find, you can quickly find a good number of other high-quality, highly relevant sources. This doesn't replace keyword searches, but it's a very quick and very effective addition.