Sometimes a company and startup do know all details about incoming traffic. The reasons may be various. In the past we could ignore this thing, because we had only one opportunity with serverfull approach. We were creating a new physical / logical servers, installing and configuring OS and making some preparations for microservices through automations.
Today, everything is changed. Huge Cloud Companies have “Function As a Service” or FaaS. Please take a look at this post from Datadog company with good statistics and explanation.
It means there is a choice between using serverfull or serverless approach nowadays!
But there are main questions which are presented below.
- How to decide whether we should to use serverless and save money?
- Will serverless approach be effective for particular company / startup project?
- How to understand that incoming traffic completely satisfies for best practices of using serverless?
Vladimir made a research based on AWS Cloud and then created a command-line tool which can analyse incoming logs from Web Servers (Apache , NGINx , etc.) and returns as a result of analysing is a verdict about using virtual servers (AWS EC2) and FaaS (AWS Lambda) from a budget point of view. The input for a tool is the cost of usage of FaaS and EC2 and can be configured.
In other words, tool can propose a new architecture in the AWS Cloud for project.
From this …
to this one …