Re: [google-appengine] Re: Search limit

2017-06-22 Thread Emlyn
I do this is a slightly more costly way Nick; I generate my new document, read any existing document from the index, compare, only put on a change. I like your memcache idea, I could possibly cache the last written version. On 22 June 2017 at 11:38, Nickolas Daskalou wrote:

Re: [google-appengine] Re: Search limit

2017-06-21 Thread Nickolas Daskalou
That's a nice way of working around the rate limits Emlyn. In your _post_put function you can also compare a hash of the last document you put in the Search API index for the Datastore entity, and only send to the Search API if the current document's hash is different. To make it fast and

Re: [google-appengine] Re: Search limit

2017-06-21 Thread Emlyn
Thanks George, I read that, that's great. I may have to go for a support level in the near future. For anyone who's interested, I was able to fix my problems with the Search API as follows: My code enqueues a huge amount of tasks very rapidly, which modify datastore objects. In the _post_put

Re: [google-appengine] Re: Search limit

2017-06-21 Thread 'George (Cloud Platform Support)' via Google App Engine
Hi Emlyn, News is not entirely bad: "Note: Although these limits are enforced by the minute, the Cloud Platform Console displays the daily totals for each. Customers with Silver, Gold, or Platinum support can request higher throughput limits by contacting their support representative." This

Re: [google-appengine] Re: Search limit

2017-06-20 Thread Emlyn
Thanks George. Not quite the news I wanted, but better to know. On 21 June 2017 at 01:13, 'George (Cloud Platform Support)' via Google App Engine wrote: > Hello Emlyn, > > The 15000 documents per minute quota is meant for the whole app. You can > see in the

[google-appengine] Re: Search limit

2017-06-20 Thread 'George (Cloud Platform Support)' via Google App Engine
Hello Emlyn, The 15000 documents per minute quota is meant for the whole app. You can see in the document you link to that the action is named "Adding documents to Indexes", so generally, not one index or per index. In the Java "Quotas" document