Sounds impossible, right? After the initial surge of cloud-native application implementations, many enterprises found that in order to ensure application performance, it’s necessary to add additional hyperscaler resources, especially during peak demands. Many enterprise policies have been to add as many hyperscaler resources as needed to avoid degradation and to maintain performance and availability. But what happens after the application peak demand subsides and the additional resource requirements dwindle?
Enterprises, many without a clear hyperscaler resource reduction plan when application demand subsides, end up paying for significant extra hyperscaler resource costs. This makes the overall cost-effectiveness and ROI for cloud implementations less than desirable. This is an issue we hear over and over again – “our cloud application performance is good, but our cloud costs are out of control.”
The good news is, help is available. Let’s take a look at the options.
Real-time Observability
Real-Time Observability monitors your cloud hyperscaler applications and provides precise 1 second metrics, full end-to-end transactions, and AI-driven context without sampling – EVER. These precise metrics and traces provide the fastest possible notification of an issue, compared to competitive observability platforms, for SREs, OPs teams, DevOps and other stakeholders so that remedial action can be initiated.
These remediation actions can be:
- Automated actions to adjust resource allocations to reduce infrastructure resource contention
- Semi-automated actions encapsulated in Runbooks or other scripts
- SmartAlert notifications that initiates manual remediation which is measured by MTTR
All of these actions are initiated by real-time observability’s precise 1 second telemetry and full end-to-end traces that provide full stack visibility. Combined these capabilities provide the fastest means for detecting issues (1 second MTTD) and providing notifications with context (3 second MTTN).
With any application issue, especially cloud-native applications, issue remediation time is of the essence.
For most application teams, any degradation in application performance, negatively impacts users, as depicted by the Prisa customer story.
Rapid detection of application issues, provided by real-time observability, enables the capability to either prevent issues from turning into incidents or when prevention isn’t possible, to rapidly triage issues when manual remediation is required.
Try the Instana Sandbox or A free, 14-day trial.
Automated Remediation Options
As described above, real-time observability enables multiple options for rapid remediation of issues and incidents based upon the cause and nature of the incident or issue. The options or automated prevention, semi-automated remediation, or manual triage.
These capabilities are provided as described below:
In addition, Instana has added a remediation option entitled the Action Catalog.
Instana Action Catalog
The Instana action catalog allows you to manage the automation that is used in Instana. You can create, edit, and delete actions, and you can use them to diagnose and remediate events. You can also apply tags to each action, and Instana can use these tags for action recommendations.
The action types supported by Instana include a 1) a documentation link to access relevant documentation to diagnose or remediate a known issue directly from an event context 2) a script on your agent by using the script action sensor, and 3) and HTTP calls to invoke webhooks or other REST API on your agent by using the HTTP action sensor. In essence, the Instana Action Catalog enables policy-driven remediation actions for your applications and infrastructure.
How to Achieve Cloud Cost Optimization
Instana and Turbonomic both also provide an additional benefit, which is the ability to achieve cloud cost optimization while preserving application performance, availability, reliability and scalability.
Instana’s role in cost optimization is in providing precise and full information to initiate preventative or semi-automated issue remediation as fast as possible and where applicable. When manual remediation is required, it reduces MTTR by as much as 50%.
With the cost of downtime being approximately $20,000 per minute, a 50% reduction in a $100,000 repair could be reduced to $50,000. This illustrates that a real-time observability is an indispensable component of cloud application cost optimization.
Turbonomic automated resource management dynamically optimizes scaled resources without compromising performance, reliability, or availability.
This makes for far more cost-effective management of applications resources that are added as applications scale but aren’t reduced when demand retracts.
Every hyperscaler resource that is allocated adds to the application infrastructure cost.
Turbonomic manages the applications resources to dynamically optimize their balance and significantly reduce the cost of cloud spend compared to implementations where resources aren’t dynamically balanced.
Try the IBM Turbonomic Sandbox or Request an IBM Turbonomic demo
