监控 AWS 中的数据库趋势

Question

I was wondering about what the best workflows/tools are for the following scenario.

Imagine you receive data from N restaurants, on a daily basis, like how many drinks, dishes of certain type, total order count etc etc, a restaurant made. All these entries go into a postgres DB, described best by the following fields {ID, datetime, restaurant, type_record, count} . Number of restaurants is in the 100's, so I need something that does not need to be updated with a CONFIG file every time a restaurant is added to the system.

Now I want to run a daily script that:

1. Runs basic queries against the DB.

2. Makes some basic calculations.

3. Catches something like number of drinks for today for restaurant X is 15% higher than its daily average`.

4. If step 3 is beyond a certain threshold, push an alert to slack or pagerduty.

The question is: with which aws service should I perform step 3?

All I can think of is to run this code on a simple lambda function . This implementation would mostly suffice but I was wondering if there are smarter/better ways to achieve this.

Details:

Latency of the query (steps 1 and 2) are not a problem, nor step 4. The main problem is how to have such a trend monitoring system on the DB that is as simple as possible (easy to maintain).

Any ideas/thoughts?

Answer 1

Either Lambda or EC2 would work. Those are the 2 compute resources AWS provides.

This type of monitoring would normally run periodically eg once per day at noon. For that type of monitoring, Lambda would be perfect, as it can be invoked only when needed.

You can also launch a Ec2 instance periodically, through a scheduled event. But there is the overhead managing the server: install software and manage an AMI.

Either would work. I suggest you try a prototype in Lambda. Lambda can simplify application development and deploy at a lower cost than developing on traditional EC2 instances.