I'm less worried about GMail. It's much easier to start a GMail competitor today (in relative, not absolute, terms) than a Chrome or Android competitor, because of network effects. For example, Fastmail is tiny (comparatively speaking), and will probably stay tiny forever, but their service works fine, and there's no major obstacles (comparatively speaking) to replacing GMail with Fastmail for any of us personally.
I think any potential competitors face many of the same pitfalls as, say, Chrome competitors do. Maybe even worse. Google slams its weight around when web standards are being designed so as to unilaterally benefit Chrome, but (at least in theory) it's a fundamentally cooperative process where everyone at least gets some sort of input to direct where each standard goes.
In GMail's case, they can just arbitrarily shut off any competitor who might be gaining steam and kill them off before they can reach critical mass and sustain themselves. Just categorize them as 'spam' and make sure to redirect their emails 100% of the time and they've won.
EDIT: just saw your edit. You're kinda right, but if Fastmail ever really starts growing then Google will take harsher actions to stymie it off. Maybe if a lot of small services start to collectively take a bigger slice of the market then they'd succeed at keeping Google at bay? I'm not so sure.
Yes, of course it is. How many search competitors do you see out there? If you ignore resellers like ddg, these is bing, kangi, and country specific ones like Baidu. Email providers are a dime a dozen - Fastmail, Outlook, Protron, .. Google it, it's a long list. Hell. I run my own. I could not even contemplate building a web indexing engine.
Not easy in any absolute sense, but very much easier in a relative sense. You're right that Google still have indirect levers to pull, to mess with independent email providers, but the distance is more and the scale is less. In comparison, Google is literally an absolute gatekeeper on Android (insofar as Google Play is the only app store that the mass market cares about, the mass market would never accept an AOSP repackage without Google Play Services, etc.), and Chrome has become the reference implementation for web standards (insofar as some increasingly large proportion of developers develop against and only ever test in Chrome).
