I share usage statistics

Recently, in one of my project it took to add some unpretentious statistics. I will not go into the details of the project itself, I will just say that it is a shareware program that costs tens of thousands of users. My goal is to know how many people a day use the Trial version of the program.

The obvious solution is to install a web server, write a small script that processes a request for a URL like http: // myproject / ontrial and then my program should launch a request to this URL.

Earlier, I already did small crafts on GAE, so there is some kind of experience, and the deployment of services attracted me. Therefore, for a long time I didn’t even think about where to locate my service, especially since it consists of 1 simple method, which essentially does nothing. What is even more gratifying is the statistics in the GAE admin panel, in which you can see which methods were twitching and how many times. Further, I cite dry usage and price statistics (many pictures)
')

Simple python handler

Porosov per second

Milliseconds per request

Bytes taken per second

Bytes sent per second

CPU usage

Number of milliseconds spent every second to process the request

Number of failures

Instance Types

My billing settings

Payment statistics

Details on payment

findings

• For an “empty” Python handler, which is requested an average of 43,000 times per day, you will have to pay about $ 30 per month.
• According to preliminary estimates, you can save up to $ 10 per month. For this, it is necessary to estimate how many on average processor hours per day are used over a quota (in my case, as you can see, this is ~ 13 hours). In the settings of the billing set a fixed number of hours that will have to pay every day, regardless of the load. But 1 fixed hour will cost $ 0.05 instead of $ 0.08.
• Taking into account the savings, I believe that in my case the use of GAE is certainly more profitable than unprofitable. Although of course you can find cheaper options.