What happened?
Starting November 1, Yandex.Maps has reduced the limits on free use of the Geocoder HTTP API from 25,000 to 1,000 requests per day. But not everyone was notified of this.
In addition, while the new limits only concern the Geocoder HTTP API, it remains paired with the map embedding service. Now, an accidental excess on the use of the geocoder cuts off your cards for a day, despite the previous limit for embedding in 25,000.
Of course, this is what happened here. And not only here . Yandex acknowledged the problem and even offered compensation .
And so about compensation begins correspondence with support.
At first, she doesn't see that we're using the API at all. But he gets his sight after our screenshot.
Then she redirects to commercial use .
Further, she does not see how we use the geocoder, which explains the lack of notification of new limits. Although our usual nature of using a geocoder is up to 100 daily with jumps up to 1,000, and several times a year - up to 25,000. The
conversation ends with a reference to the agreement, where it is written that Yandex is not obliged to you - everything should work this way.
The correspondence took 3 days, of which the first day the cards did not work. no one began to compensate for anything even during the trial.
But the legs of this began to grow a little over a year ago. "Why did this article appear" reminds something - yes? :)
Versioning by Schrödinger et al.
Before the limits, versioning was the main problem.
So if you indicate the version number, say, 1.65, which one do you expect to get in the end?
I thought it was 1.65. But no, Yandex.Maps is firmly convinced that it can be 1.72, 1.75, etc. Well, at least it only grows. But it is not exactly.
When asked where they found out that this is a good practice, they answered with a link to their own documentation about versioning. Cleverly - yes? :)
Thus, you receive bugs that you did not expect. You need to edit them immediately - again, no one notified you, and there is nowhere to roll back.
In the beginning, I rationalized this approach by thinking about quick feedback and making edits. Therefore, he reported everything he found. But again I did not guess. As I understand you, @ ReDev1.
Support could react for years. And even fixed bugs can be returned.
Those. you yourself must track and know all the changes, even those inside the code, and make changes in accordance with Yandex's work plan.
So it turns out that you work either with a lack of responsive map styles, or with non-working tile loading for object maps, or with an undercache of background tiles, etc.
I won't waste time on documentation and code architecture, which often resembles an Italian dish.
It's just business, nothing personal.
This is a cynical approach. And the business turns out to be bad when marketers who are failing put money above their reputation.
Another confirmation of this is the recent breakdown of a big deal with Yandex. Still, someone takes into account the mood of users.
We build our projects in a different way.
Prior to that, the Yandex.Maps API had been used for more than 7 years due to their coverage and convenient limits. But now, given the nature of use, new limits, pricing and quality of support, it makes no sense. It is much cheaper, more convenient and more flexible to use your own solutions based on other open source and free products.
Moreover, as it is, I had to write my own wrappers to solve the problems of Yandex.Maps services. For example, the same geocoder is trained on user input statistics. Because of this, it often happens that the exact address is not geo-coded, but, for example, without a street view, it is.
I think many other mid-sized projects will do the same. As a result, only small projects that use the most basic functionality will remain on Maps, and commercial users, who will receive more and more missed bugs. As a result, their outflow will begin.
Another interesting consequence of this whole story is that I stopped perceiving Yandex employees as serious specialists. And you have to divide, at least by 10, everything they say, even if there is a desire to listen to them.
Perhaps funny.
Somehow Yandex in Real Estate acquired functionality that was very similar to what was the unique proposal of our project.
Then the inner paranoid thought about the coincidence of coincidence. Because 5 years earlier, we had begun to solve the problem of finding housing, taking into account the quality of life. In particular, taking into account ecology, and using interactive maps. And this is still a very niche service.
PS
There is something similar about Google. I am thinking of writing something too, if you finish writing it.
Who knows if there is a plugin for leaflet so that you can use it with a backend under the Yandex.Maps API? If anything, I have already started writing a similar solution.