roman's comment was about slowdown when Joplin hit somewhere around 1000 tags, and about the importance of RAM because a lot gets loaded. The only limit I can see quoted is the SQLite limit which is so far beyond what a PC/Mac would be able to directly support it's essentially meaningless.
My reading of the responses here suggests that there is no practical limit to the number of notes other than that imposed by the database used which has a huge and built-in limit that you're not going to hit. Also performance will be affected by the number of tags and notebooks and the size of any individual note you're using at the time.
Machine performance is as much dark art as science as work has taught me over the years. These days it will be affected by web browsers (particularly but not exclusively Chrome), all other apps loaded whether they're running or not and the huge number of code fragments loaded by the operating system. More RAM is always good if you're dealing with a lot of data and SSDs will make the machine respond faster. These days more applications have soft limits than useful hard ones in that they gradually lose performance rather than just plain stop working at all.
If you have a need for indefinitely large amounts of well indexed data, then Joplin may not be the best solution. In fact you may be looking more at an enterprise-capable knowledge management solution which is a whole new world. The ones I've experienced are solutions for a different order of problem than Joplin (no insult to Joplin here) and the cost of ownership is high as a result.