We run 2 apps on Google AppEngine, one in Python, and one in Java.
The Python app faces the internet and serves browser users or API calls from our smartphone app.
The Python app often communicates with the Java app, which offers workflow engine services.
The Java app is never accessed directly from the internet. This is similar to a service oriented architecture.
Our Java app is configured with 1 resident instance.
However, we often see that when the Python app sends a request to the Java app (which is doing nothing at that moment), that a new Java instance is spawned, instead of having the resident instance serve the request. This way, latencies up to 10 seconds were discovered, whereas the request handling itself takes less than 1 second.
Could it be possible that Google considers a request from one appengine app to another one, as a sort of backend workload, for which throughput is more important than latency? Are there different policies for requests coming from the internet versus requests coming from Google internally (defered methods, cron, ... maybe inter-appengine requests also resorts in this category, which is maybe not always correct).
Google recommended me to run the Java app as a backend of the Python app, and in fact have only 1 single Appengine app, instead of 2.
Some information in this thread:
https://groups.google.com/forum/?fromgroups=#!topic/google-appengine/8O7K3cFzBbY
Any advice on how to force appengine resident instances handle incoming requests, rather than spawn and warm up new instances, which might be good for throughput, but bad for latency?
Have tried to add <threadsafe>true</threadsafe> element to appengine-web.xml on your java app?
Related
I have a relatively small Java app, which I'd like to move over to the Google App Engine. It runs in the console, with no user input needed after the initial startup. I researched a bit on how to deploy it, but all tutorials seem to focus on Java web apps, when I don't really need that. Is it possible to deploy my app if it's not a web app?
App Engine is probably the wrong GCP platform for you - you'd probably be better served just deploying your jar directly onto a Google Compute Engine node. GAE is pretty explicitly oriented around web applications and you'd need to do a bunch of configuration in order to have it work for your use case.
Does your non-web Java app handle web requests? If not, it seems difficult to imagine that you would be able to reach your deployed app and use it for any purpose, once deployed. Your Java app should be able to handle requests, to make deployment worthwhile, and the deployed app useful.
You may find out about how your app should handle requests by reading the How Requests are Handled documentation page.
Can google app engine be used running a web crawler?
I have a crawler process that starts with a specific id (or ASIN number) of an amazon web store product. The process then uses amazon product advertising API to get similar products. After the similar products are fetched, their ASIN numbers are used to call the API for other similar products. This process is suppose to get the information of about 25,000 products which roughly takes 3-6 hours.
Now my client wants to run multiple instances of this process with different initial ASIN numbers. I gave him a proposal on how the whole system can be made using google compute engines, but he is eager to use google app engine. He thinks that app engine will somehow automatically make crawler instances.
Since google app engine is used for backends and APIs, I have concerns that deploying such a crawler system on app engine would have its limitations.
Yes. You could fetch the base URL, scrape the data, get the urls for related products, then spin up taskqueues to fetch each subsequent url. Rinse and repeat.
Im creating a Node.js website that probably won't have loads of traffic, and was looking into cheap solutions to host the site. Came across Google cloud services offering free usage for their services with limits. A f1-mirco is more than enough for my needs, but I will happily pay for some usage if it goes over by any chance.
I wanted to setup a linux centOS 7 on GCE (which I already did), and run my application and REST API on it. Now here comes the problem.
I tried to use Google's datastore service, but it sprung an app engine instance and without it datastore won't work.
Is datastore entirely relying on app engine to function?? In the docs, it said if you use any of the client API, it requires app engine. What can I do to not use the client api and query data then? Don't want to use the app engine at the moment or datastore is just not for me then?
Thanks for any help!
Some of the underlying infrastructure of Cloud Datastore and App Engine are still tied together for creation, etc. So while creating an Cloud Datastore database also defines an App Engine instance for the project, it doesn't require you to use it. You don't get charged for App Engine either, unless you decide to deploy an App using it.
You should be totally fine use the Google Cloud Node client library on the f1 micro instance.
I am going to build an API for my mobile app using GAE. I am looking for a way to separate the whole application into separate services, trying out the microservice architecture.
The problem is there seems to be no information about any VPN or private network between GAE apps. Therefore, based on my understanding, when one of the GAE apps sends HTTP requests to other GAE apps in same account, its traffic will be treated as Internet traffic and therefore I will be charged for outbound bandwidth.
Am I correct?
Yes.
App Engine applications/Cloud Platform projects are isolated from each other by default, even if they were created using the same Google account.
Consider looking into Modules (link for Java) to implement individual services of your application.
I'm developing a Java app in GAE, which offers an API through Google Cloud Endpoints.
Basically it receives requests in the endpoints and uses a number of web services from different providers, stores some data and returns some data through the endpoints...
I understand that my app is conceptually a backend, because it doesn't provide any web page, but only the endpoints, don't you think so?
But there's no way to create only a backend, without being associated to any frontend app, is there? At least Google Plugin for Eclipse only allow you to "Generate App Engine Backend", from an existing app, and moreover this app must be an Android project...
I'm using it as a frontend and there's no problem, but apart from the conceptual issue, I've read that backends are kind of optimized to be backends, with more memory and CPU...
I think you're just confused because the Cloud Endpoints documentation uses the word 'backend' to refer to the entire cloud-hosted server implementation. It doesn't specifically refer to the use of GAE backend instances. Endpoint requests can be served by frontend or backend instances, based on how you set them up and the url being accessed.
From the App Ending docs:
"When an application is called to serve a web request, it must issue a response within 60 seconds"
"App Engine Backends are instances of your application that are exempt from request deadlines and have access to more memory (up to 1GB) and CPU (up to 4.8GHz) than normal instances."
So unless you're requests are doing something crazy, you don't need to use a backend. In the google-plugin-for-eclipse, "generate appengine backend" is talking about creating a backend for your android app... a server for your android app to contact (in this case your android app is the frontend and you're appengine app is the backend). In the example app you can remove the web side (index.html) to the appengine application and you'll have no web frontend. Index.html is using the gapi javascript library to make endpoints calls to your appengine service.