Amazon Web Services (AWS) this week expanded its campaign to become the primary destination for deploying containers by making available Amazon CloudWatch Container Insights in preview.
Announced at its AWS Summit event in New York, Amazon CloudWatch Container Insights is the second container service announcement made by AWS this week. AWS earlier announced the general availability of AWS Fluent Bit, which enables container logs to be routed to the AWS CloudWatch management service or through the AWS Kinesis Firehose streaming data service to allow them to be stored in the Amazon S3, Amazon Elasticsearch or Amazon Redshift services.
CloudWatch Container Insights makes it possible to monitor and troubleshoot containerized applications and microservices running on all the major AWS services. It provides access to dashboards that automatically discover and summarize operational data from container environments, monitor and analyze granular container performance metrics to determine resource use, and provides access to tools for troubleshooting containerized applications and microservices.
AWS CTO Werner Vogels told conference attendees that containers have become critically important in the age of the cloud because they provide the ideal environment for scaling applications up and down. Rather than being forced to rely on a single pipeline to build a monolithic application, containers make it feasible to employ multiple pipelines to build applications based on microservices-based architectures much faster, he says. As it happens, Vogels notes, the AWS Elastic Container Service has the deepest integration with the rest of AWS’ cloud services and that fast-food giant McDonald’s is using the service to drive a home-delivery application capable of processing more than 20,000 orders per second.
Vogels also observed during his keynote that containers and the Kubernetes platform they are deployed on are a pain to manage. Developers not only find themselves having to build and deploy containers using low-level tools, IT operations teams are confronted with a complex Kubernetes platform that is difficult to provision and update dynamically. As part of an ongoing effort to automate DevOps processes end to end, AWS is committed to eliminating the friction surrounding the management of containers and the platforms they run on, says Vogels, adding development teams should enjoy working on their applications rather than having to manage Kubernetes.
AWS is also making extensive use of containers to build a variety of serverless computing services that developers can invoke on-demand using functions. Rather than having to build a separate set of containers for each task, serverless computing frameworks allow developers to dynamically invoke, for example, an AWS EventBridge service to route data between applications when needed.
The challenge many organizations will face, however, is that as they invoke more AWS services, the more difficult it becomes to move an application to another platform because of all the proprietary hooks and application programming interfaces exposed by AWS. It’s also challenging to extend a common set of DevOps process across multiple clouds when relying on the toolchain provided by AWS.
Nevertheless, AWS services are already home to billions of containers running in a production environment that, in terms of size, continues to expand faster than any other.