eBay
With eBay, you're in an essentially laissez faire market. This means the onus is on you, the buyer, to understand both what you want, and what you have a reasonable expectation of paying for a given product. As a buyer, it is not a very good avenue for the complete novice who doesn't understand the difference between a resistor and a piece of wire, mainly for the reason that they are looking for keywords or buzzwords that align with what they think they need.
"Don't buy from china," - the second hand Rubidium Standards almost exclusively ship from China, and are recycled. As a buyer who knows, it's a hedging of bets. Do they have good reputations from others who bought the same product? Is the price within range of the other market offerings? It's a good bet, but not a guaranteed bet. Is the risk vs. reward at (now) $200 a pop worth it? Semiconductors - same thing. Do they source high volumes of specific parts, or is it a complete hodge-podge of random parts? "Everything from China is black market!" - One of the biggest "games" played by Chinese ODMs is to use a lead customer's bill of materials to get pricing down, lump in a particular part with other customers, and get even higher volume pricing to improve their margin. If their lead customer is buying 250k pieces, and their second-through-fourth customers are buying 100k pieces each for a total of 550k pieces, they may push the part supplier to sell their $0.10 LDO at 500kpieces, to sell them at $0.05 at a 750k pc order, while charging their own customers $0.15/ea. They then turn around and resell the difference in stock (200k pc) for $0.25 each, while you and I would pay $1 from Mouser. If they sell all 200k pieces, they've made $40k for what is essentially consolidating initial orders.
What usually comes up, is that a newbie sees "74 series logic grab bag: 250 pieces" for $50, buys it, and most of them are dead because they're either factory floor sweepings, re-worked part pulls from recycling, or some other bottom-of-the-barrel operation that might have paid $1 for all the scrap. The experienced buyer sees that they don't have any particular rhyme or reason behind what they sell; and it all appears to be scrap. They know to stay away.
Manufacturers
It depends.
Almost every engineer will have a handful of companies on their own blacklist. You will often find that many of them are highly reputable companies, too. At some point, the designer will have chosen to use a brand's part, either it looked good, or worked well in early prototypes, or something along those lines. Then some timed event, either come purchasing, production, or long after its been in the field, there is some failure. Sometimes they didn't read the datasheet and missed something. Sometimes the silicon lot has a terrible defect in it and an entire lot snuck by probe or test because of an idiotic hardware fault. Sometimes there is a design oversight and the part is put into a particularly stressed condition. Whatever it is, this part is now a problem for the engineer and it comes down to the experience they have with the vendor as they work to resolve it. If the vendor is good, and they get it fixed, they react one way. If the vendor is good, and they cannot fix it, they react another. If the vendor ignores them because they're a Tier 3 or some other name of a small guy, and they get no where on useless forums or similar places, they feel scorned and never use another part from that vendor. Monetary impact has a huge role in it, too.
In a recent design I did, I needed a simple buck-mode switcher. I used a part from a company with 3 letters abbreviation, used their calculator tool, got some numbers, and did a prototype. It essentially smoked. It "powered up" to 1.02V unloaded, and nothing under load and the part got very hot (eventually reflowing its own solder). Everything within the calculator's specs, and same parts as it commanded. Nothing. I could have monkeyed with it for ages, but I decided to discard the supply, verify everything else (which worked just fine) and did a redesign with another switcher made by another 3-letter-abbreviated company who happens to make a popular spice tool. Boom, another turn of prototype PCBs, which included a few minor changes to the design in addition to this new Switcher.
Nothing.
It would not power up at all. All of the nodes had the right supplies. Everything matched the datasheet (canned circuit right off of page 1). Just... nothing. I shot this company a tech support inquiry on their webpage and on their twitter. After 3 days, nothing. I moved on to another popular vendor, who only has two initials. The lowest cost solution of them all. The fewest parts of them all, too. I simply take a hit for inductor size (lower switching freq) and lower input rail (18V instead of 20 or 24V). A third round of boards. Power it up... and...
It works. Perfectly.
So now I personally have a preferred vendor of switching supplies, because of whom I had success with. I will then bet you that I can find an engineer who had the exact opposite experience I had, and exclusively uses another vendor. But note, I don't automatically exclude all of the parts made by other companies; but specific lines of products which seem to have inferior support models.
It'll be the same thing with passives, and discretes.
The fundamental value you get with a distributor like Mouser or Digikey, is that if a part fails its advertised claim, you can call them up and they'll ship you a replacement, and likely chase down the manufacturer to deal with the bad material.