Skip to main content

The hidden costs of serverless and microservices

Container-based solutions were one of the best ways to reduce running costs and improve the quality attributes of a product five years ago. Nowadays, IT departments are complaining about the running cost of containers and pushing serverless as the next step to reduce the running costs.

It is important to identify why a container-based solution is seen as expensive and what are the cost vectors before saying that a serverless approach using containers is less expensive. 


A container-based solution already provides us with a degree of flexibility, allowing us to spin up & down the number of instances of services. Microsoft Azure, like other cloud vendors, provides the flexibility to run our container-based solution in dedicated or shared clusters. At this stage, I would like to mention a few available options:

(1) Dedicated cluster: Azure Kubernetes Services (cost are driven by the cluster size)

(2) Serverless approach: Azure Container App (cost are driven by the computation usage)

(3) Hypervisor isolation: Azure Container Instance (cost are driven by the computation usage)

(4) Dedicated cluster with 'overflow capacity: Azure Kubernetes Services together with Azure Container Instances

Additionally, we have Azure Functions where we can run our containers in a serverless approach when we build an event-driven solution. 

Resource reservation

The technology is already available to build solutions that are scalable and can go from 1 to 100 instances in a few seconds. As long as we have the resource available where we can spin up our instances. 

To ensure that we have the computation resources available, we start to reserve computation resources and pay for resources we don't actually need at a specific moment. When going with a serverless approach, we need to ensure that we don't end-up with the same situation, running our serverless services in a dedicated cluster to ensure that we have enough resources in the case of a spike. Otherwise, from the cost point of view, we would be in the same situation.

Compliance Regulations

Technology is one thing. Law and regulations are more important and more powerful. The current technology is giving us the ability to run our container-based solution using a pay-per-usage approach. Azure Container Instances allow us to run our services isolated from other customers that use the same cluster using hypervisor isolation. The cost model is pay per usage, and it's very scalable.

When you work under HIPAA or PCI-DSS, sharing computation resources is not allowed by regulations. You are forced to use a dedicated cluster that runs your own payload and is isolated at the network and computation layer. 

You also need the ability to monitor all the traffic, making a dedicated Kubernetes cluster like Azure Kubernetes Service a good option. As you expect, dedicated clusters are more expensive than Azure Container Instances or Azure Container App. Even if you don't need that computation power, you would need to pay for it.

In regulated industries, where computation and network isolation are mandatory, you would need dedicated resources, serverless or not, they are more expensive in comparison with other tiers or services. For example, a vCPU core for Azure Functions Premium tier would cost you ~$124/month. Yes, it is offering a lot more than the Consumption tier, but you pay per core - regardless if you use or not the computation power.

Final thoughts 

Serverless and event-driven solutions are one of the best ways how we can improve the quality attributes of our systems. In terms of costs, factors like compliance regulations and resource reservation can force us to go with 'dedicated' tiers, where serverless or not, the running cost will be more expensive. We should remember this and set clear expectations when discussing microservices and serverless approaches. 

Comments

Popular posts from this blog

Windows Docker Containers can make WIN32 API calls, use COM and ASP.NET WebForms

After the last post , I received two interesting questions related to Docker and Windows. People were interested if we do Win32 API calls from a Docker container and if there is support for COM. WIN32 Support To test calls to WIN32 API, let’s try to populate SYSTEM_INFO class. [StructLayout(LayoutKind.Sequential)] public struct SYSTEM_INFO { public uint dwOemId; public uint dwPageSize; public uint lpMinimumApplicationAddress; public uint lpMaximumApplicationAddress; public uint dwActiveProcessorMask; public uint dwNumberOfProcessors; public uint dwProcessorType; public uint dwAllocationGranularity; public uint dwProcessorLevel; public uint dwProcessorRevision; } ... [DllImport("kernel32")] static extern void GetSystemInfo(ref SYSTEM_INFO pSI); ... SYSTEM_INFO pSI = new SYSTEM_INFO(

ADO.NET provider with invariant name 'System.Data.SqlClient' could not be loaded

Today blog post will be started with the following error when running DB tests on the CI machine: threw exception: System.InvalidOperationException: The Entity Framework provider type 'System.Data.Entity.SqlServer.SqlProviderServices, EntityFramework.SqlServer' registered in the application config file for the ADO.NET provider with invariant name 'System.Data.SqlClient' could not be loaded. Make sure that the assembly-qualified name is used and that the assembly is available to the running application. See http://go.microsoft.com/fwlink/?LinkId=260882 for more information. at System.Data.Entity.Infrastructure.DependencyResolution.ProviderServicesFactory.GetInstance(String providerTypeName, String providerInvariantName) This error happened only on the Continuous Integration machine. On the devs machines, everything has fine. The classic problem – on my machine it’s working. The CI has the following configuration: TeamCity .NET 4.51 EF 6.0.2 VS2013 It see

Navigating Cloud Strategy after Azure Central US Region Outage

 Looking back, July 19, 2024, was challenging for customers using Microsoft Azure or Windows machines. Two major outages affected customers using CrowdStrike Falcon or Microsoft Azure computation resources in the Central US. These two outages affected many people and put many businesses on pause for a few hours or even days. The overlap of these two issues was a nightmare for travellers. In addition to blue screens in the airport terminals, they could not get additional information from the airport website, airline personnel, or the support line because they were affected by the outage in the Central US region or the CrowdStrike outage.   But what happened in reality? A faulty CrowdStrike update affected Windows computers globally, from airports and healthcare to small businesses, affecting over 8.5m computers. Even if the Falson Sensor software defect was identified and a fix deployed shortly after, the recovery took longer. In parallel with CrowdStrike, Microsoft provided a too