Good day,
We're relatively new clients to using your service and noticed our server costs are quite high. The App Engine appears to be charging around $4-5 per day, and we are only using it for internal purposes with 4 employees logging on daily. The total cost monthly is around $110 per month for the App Engine. I just wanted to ask if this was normal for companies our size, or if there was something we could do to lower the cost. The frontend client makes less than 1 request per second to the backend api server. The memory usage ranges between 136-490 mb. There are about 2-3 F2 instances on average during work hours. Are we missing any other variables that affect the cost? It seems strange that we are using 2-3 F2 instances on average when our app is extremely simple: Users log in, they are presented with basic data from our database, and every few minutes they click Submit and make a call to the database. Where could this high instance count be coming from?
Thank you for your time,
Jonathan Tan
Just in terms of the instance costs, ~$110/month is about what you would pay for 2 F2 instances running 24/7 (obviously, this varies depending on your region), which certainly does not seem to align with the usage you describe. Running 2 F2 instances for 8-10hrs/day, only on work days, should be roughly $30-$50/month (depending on region) for instance costs (assumes 8-10hr workday, 20 workdays/month). Also, 2-3 F2 instances for 4 employees seems very high, given the low requests per second, etc., so there may be a fundamental issue with how the app is designed.
The most obvious thing to check is if you have set either min_instances or min_idle_instances in your app.yaml, as this will keep instances running whether they are being used or not.
Just a note, all of this only applies if the ~$110/month is instance hour costs. If it is other usage, like network egress, that all comes down to how you have written your app.