Building a containerized software system is easy. Deploy it to Kubernetes, and your product is suddenly online, highly available, and delivering user value. But how do you manage it over time? What about a large fleet of such deployments and services? Is your software stack still delivering the same value, cost benefit, and following industry best practices?
It’s easy to let systems slip out-of-date, or tweak replica counts, port settings, or configuration values and suddenly a system is out of compliance, wasting resources, or worse, falls offline.
This Shoreline automation Op Pack frequently scans all the resources in Kubernetes, comparing them to the source-controlled YAML files, config store keys, and industry best practices. If a resource falls out-of-compliance, an alert is generated in Shoreline and optionally in an external system. A user can quickly confirm restoring the system, and the corrected YAML diff is put in place.
Special care is taken with horizontal autoscalers to insure we’re not looking at the deployment’s, StatefulSet’s or DaemonSet’s replica count, but instead looking at the bounds for the horizontal autoscaler, and ensuring only the necessary replicas are in place.
