Summary:

Flexera is one of the leading IT Asset Management companies, giving their users a complete view of their IT infrastructure, including their cloud inventory. It was a natural extension of their portfolio to provide reporting on cloud costs in addition to the cost of SaaS and on-premise IT applications. However, the depth that Flexera has applied to IT asset management (ITAM) and governance has only partially made its way into their cloud support. They lack any support for Kubernetes or spot instance orchestration, which means their customers are missing out on more cost savings.

Anyone that deploys Kubernetes in their cloud infrastructure knows that shared cluster costs can quickly spiral out of control if they aren’t proactively managed, and that’s where Harness wins. We give our customers rich, granular insights into their Kubernetes costs, provide right-sizing recommendations at all levels of the cluster, and proactively manage cluster scaling and costs with Cluster Orchestrator, saving our customers anywhere from 70-90% on their Kubernetes cluster costs.

Anomaly Detection: 

Misconfigured scripts, test clusters left running, run away Lambda functions…all of these things can cause your cloud bill to spiral out of control unless you have intelligent cost anomaly monitoring and alerting in place. Not having any automated cost monitoring in place is bad, but having too many distractions in cost anomaly alerting can cause important alerts to go unnoticed. 

Flexera’s cost anomaly alerts is a static reporting module that cloud administrators have to manually configure to receive alerts (brush up on your standard deviations). The inputs as to what constitutes an anomaly are a “best guess” effort from the administrator attempting to find a “one-size-fits-most” anomaly threshold. It’s also a global alerting policy that can’t be customized to alert only affected users because they don’t support custom tags in the reports. Everyone gets only email alerts, so no support for Slack or other integrations. For large enterprises, this will be a lot of alerts that lack the data needed for prioritization.

Harness cost anomaly detection leverages artificial intelligence and machine learning (AI/ML) to dynamically detect and alert on cost anomalies as they occur, using a variety of statistical methods to find true anomalies and filter out the noise. We report anomalies to the affected users by Cost Perspective, so only the folks that need to take action are getting alerted. We deliver actionable alerts without a lot of noise.

Kubernetes Support:

Modern SaaS infrastructure relies on containers and Kubernetes for easy scalability and flexible deployment models. However, shared cost allocation is a challenge unless you have the tools in place to dig into costs by workload/node/pod/label, etc. Harness CCM provides detailed cost analysis for Kubernetes clusters at a granular level that can be tailored for your business. We also deliver automated cluster cost management features – such as Cluster Orchestrator for EKS – that can save our customers up to 90% on cloud costs. Flexera has no support for Kubernetes at all.

Idle Resource Management:

One of the biggest sources of cloud waste is idle cloud resources, especially in test/dev environments. Your engineering teams don’t work 24/7, so why would you want their cloud instances to? Shutting down idle cloud resources while they aren’t in use can deliver major cost savings, but infrastructure teams struggle to automate this to reduce friction with engineering productivity.

Flexera does provide schedule based instance shut down, but it’s little better than a manually maintained Lambda script. While this can save on cloud costs, it leaves a lot of cost savings on the table, especially for test resources that are not used consistently during working hours. Any deviations needed from the predetermined schedule need manual administrative attention to execute.

Harness Cloud Autostopping automatically detects and halts idle cloud resources when they aren’t in use, and brings them back up again once an application or engineer accesses them. Our customers configure the Cloud Autostopping policy once, and we do the rest, freeing up DevOps teams to focus on more value added activities than managing manual shut-down schedules.  Harness customers save up to 80% on the cloud costs with Cloud Autostopping.