Fair use of api (max number of requests, max number of locations in parameter) about open-elevation HOT 3 CLOSED

Hi!

Sorry for the long delay, but work is hectic, and that leaves little time for side-projects...

There is currently no limit, but I can tell you that we are currently averaging about 1-2 requests per second. If you could keep your requests below that, the server might hang-on and "survive".

I would very much like to improve our current system, but the maintenance and hosting costs are currently fully covered by me, and sadly I am not rich...

I am considering creating a system with API keys so that it is easier to manually throttle and/or warn clients. I understand that it might break people already using the API, but with the popularity that it is slowly (yet surely) gaining, I am afraid that what you mention here might happen: that someone might accidentally render the service unusable for others. An API key would help with that.
(EDIT: we could keep the current API, but throttle it, thus maintaining backwards compatibility)

I would also suggest that you make requests which are geographically close to one another for maximum performance. The server caches the most recent tiles.

Regarding max client header buffer, we are currently forcing request lines to be at most 4096 bytes, but there are plans to extend this value and to provide a POST api. I just need to find more free time...though any PR's are welcome!

If you have any more questions, please do say!

from open-elevation.

Slight update on this: Due to high amounts of traffic, a rate limit has been imposed now: 1 req/sec for each IP.

I have been looking into creating a commercial version of the platform so that I can improve the public version. As usage increases, the overall performance for everyone decreases, and that is not my goal.

from open-elevation.

Hi.

For the better part of the last 2 years, the API has been completely overloaded (sadly, very little comes in the way of donations, and so I can't really pay for a more beefed up machine).

As of two days ago, though, I have done some tweaks to the backend and the API should be handling the workload much better. Give it a try again!

I will also be pushing a series of updates to the repository and dockerhub, to modernize the project and make it work again.

I will close this issue because I believe that the public API is now working, but feel free to reopen it if it is not. If you have any other issues, please open a separate issue for that.

from open-elevation.