tomhughes left a comment (openstreetmap/openstreetmap-website#6572)
The problem is no matter what timeouts you put on individual parts of a request
you can't guarantee a hard upper limit on the overall request, at least not
without significant effort to evaluate the worst case on a request by request
basis - the whole point of the current system was to try and create a backstop
that guaranteed a universal limit for all requests.
In many cases the database query will be the main part of the request so
reducing the query timeout from where I've set it would amount to tightening up
the current limits for API requests.
As things stand the query timeout seems to be working and the big spikes in
running queries have stopped since I added it on the 28th:
<img width="3200" height="1800" alt="Image"
src="https://github.com/user-attachments/assets/fddb62a6-8b0f-4ee1-8c65-a1dbc415c51b";
/>
--
Reply to this email directly or view it on GitHub:
https://github.com/openstreetmap/openstreetmap-website/issues/6572#issuecomment-3593712018
You are receiving this because you are subscribed to this thread.
Message ID:
<openstreetmap/openstreetmap-website/issues/6572/[email protected]>
_______________________________________________
rails-dev mailing list
[email protected]
https://lists.openstreetmap.org/listinfo/rails-dev
