REMINDER
Ran out of API calls per minute, but no one's using it...
-
I just got the notice for a site that's not even in production.
Seems it got 140 hits in one go, but it's only bound to pages and the page-stats shows no hits at all.
Is there any way to block public access to API and only allow bindings?
-
What is even weirder is that it's all hits from today, 141 of them - but GET requests are supposedly cached, which doesn't seem to be the case.
Something seems very fishy here.
-
Is your API call used for a dynamic data source page?
-
@Wassim said in Ran out of API calls per minute, but no one's using it...:
Is your API call used for a dynamic data source page?
Yes
-
Then it's not cached, and I guess your website has a lot of dynamic pages indexed that have been crawled by Google or any other bot?
-
If it's not cached, then I'll just have to abandon Appdrag.
The whole point was that it could handle traffic-spikes and that API was not called on every page-view.
-
Joseph wrote this:
"2) We do have a caching layer for ALL GET requests to your API, 30 seconds by default So if in 30 seconds timespan you get 1000 API calls to a GET endpoint, this is accounted as only 1 API call! not 1000, because the other 999 are served by cache"So this is what I have expected.
-
If that's not the case, then tell me now so I'm not wasting any more time.
-
There is a 30 seconds cache. What I'm trying to explain is that the cache is about a specific url.
So if you have a dynamic data source on a page this is what will happen during 30 seconds:
- yourpage.html?product1 => no cache first request
- yourpage.html?product1 => 30s cache
- yourpage.html?product2 => no cache new product
- yourpage.html?product2 => 30s cache
etc..
So if your database contains a lot of rows and your website has been crawled, that will happen. -
@Wassim
I understand and that's what I expect.The free limit is very low so I'll up the cache to 5 minutes and see if that helps.
-
@ThomasD You won't be able to go to production with our FREE tier of cloudbackend ...
this is for evaluation and testing only ... all our other plans haver larger API calls per minute + burst (5 times your limit)Also the limit is just for a timespan of 5 min when you are over quota
-
@Joseph-Benguira Ah, not sure how I missed that FREE tier is not for production, is that a recent change or always the case?
And yes, I see I'm unable to change the cache time..
-
@Wassim said in Ran out of API calls per minute, but no one's using it...:
etc..
So if your database contains a lot of rows and your website has been crawled, that will happen.How do the database relate to cache?
I thought cache applied to API endpoints and not database?
-
Cache is based on the parameters:
If you make an API call to- getMyProduct?item=chocolate => you expect to have chocolate
- getMyProduct?item=lolipop => you expect to have lolipop, not chocolate because someone before you cached chocolate no?
About your question on production and free tier, it's an advice but since you have many rows/pages, you will be burst every time your website is crawled and for your case to continue on free tier will end up with crawling issues and not the best SEO you can have as some pages won't be able to be crawled as your quota will be exceeded.
-
Thanks it all makes sense now.
I was about to use the Starter Tier for our use-cases, since I think that's the minimum required and also use 1 minute cache.
I'll probably make a calculator for back-end, so I don't lock myself into unnecessary billings, maybe it could be useful for the community too.
