Over the past few years, we have architected and deployed an ideal observability in healthtech and healthcare organizations, gaining valuable insights along the way. Drawing from our experiences with Innovaccer, Tata 1mg, and others, we’ve identified several key lessons to share with other healthtech companies, helping them navigate the complexities of observability.

1. Fragmentation Hinders Efficiency

Both Innovaccer and Tata 1mg began their journeys with fragmented observability stacks, which significantly slowed their incident response times. Innovaccer relied on a combination of Thanos, Prometheus, Grafana, and NewRelic, resulting in scattered data and extended Mean Time to Detect and Resolution (MTTD/R). Tata 1mg faced similar challenges with its observability spread across Elastic APM, Fluentd, Kibana, and AWS CloudWatch. This fragmentation not only drove up costs but also limited their insights to identify root causes of performance issues.

Transitioning to Kloudfuse allowed both organizations to consolidate their observability data into a single platform, enabling faster issue detection and resolution.

2. Safeguarding Sensitive Data with Private Deployments

In healthcare, data privacy and security are non-negotiable. While metrics and traces typically don't contain Protected Health Information (PHI), logs can. This concern has been prominent among all healthcare organizations we have worked with.

Kloudfuse’s VPC deployment keeps data within their environment and enables them to control it fully. As GE Healthcare noted, implementing “Kloudfuse's unified observability data lake secures all metrics, events, logs, and traces (MELT) data, ensuring compliance with regulatory standards while centralizing observability efforts.” In addition, all the transfer associated with the observability data is optimized and made secure by internal mechanisms using private links in the cloud. The data in transit and rest is completely secure using the customer adopted encryption mechanisms. 

Some of our healthtech clients also choose to deploy VPCs within each of their client environments, providing further flexibility and control. These organizations scrub PHI at the source to minimize risk, emphasizing their desire to de-risk unnecessary data movement—especially concerning sensitive information related to HIPAA. Combined with robust security measures, including RBAC, 2FA, SOC 2 Type II compliance, penetration tests, and data encryption, this approach offers comprehensive protection.

Additionally, our clients have recognized the value of Kloudfuse’s unified observability data lake, which provides a centralized platform capable of protecting sensitive data in one place. This eliminates the risks associated with data replication across multiple tools, which can lead to errors and security vulnerabilities.

3. Data Isolation Based on Internal Consumers

To achieve data segregation for their internal customers, Innovaccer used the attributes of observability data related to their internal consumers and leveraged the capabilities of the kloudfuse platform to isolate data based on those attribute values, such as customer id or name. This approach clearly isolates sensitive data  per customer, so that specific teams in the organization can access it easily in the user interface.

4. Self-Service Capabilities Foster Agility and Adoption

Organizations like Innovaccer are increasingly embracing self-service observability. They aim to shift from an enablement model to a self-service paradigm. Kloudfuse’s self-service capabilities have empowered teams across Innovaccer to create their own dashboards and alerts, fostering a culture of collaboration and empowerment. Currently, over 500 engineers, DevOps professionals, and Site Reliability Engineers (SREs) are leveraging Kloudfuse at Innovaccer, while Tata 1mg has approximately 300 users.

These self-service features, including the ability to create custom dashboards, monitors, and alerts, combined with Role-Based Access Control (RBAC), have enabled teams to manage observability independently. This approach has significantly reduced wait times for access and alleviated the overhead of relying on a centralized team for observability management.

Similarly, Graphite Health’s adoption of Kloudfuse has eliminated manual monitoring overhead, allowing their operations team to concentrate on strategic initiatives. This shift not only lightened their workload but also sped up deployment and made it easier to respond to customer needs.

5. Managing Cost is Key as Growth Accelerates

Organizations are keenly aware that data volume growth shouldn't translate to spiraling Observability costs. Many have also learned from painful experiences with overage and egress fees when transferring data to observability vendors.

To address these challenges, healthtech companies are rethinking their observability stacks with cost management as a priority. Kloudfuse provides fixed pricing, along with data transformation and deduplication capabilities, which can significantly reduce data storage and data analysis costs. Additionally, by keeping data within the customer’s environment, Kloudfuse eliminates egress fees, avoiding the costs associated with transferring data across the network to a hosted vendor.

We are delighted to hear that Tata 1mg experienced a remarkable 40% decrease in costs after implementing Kloudfuse, even while doubling their data volume.

6. The Hidden Costs of Open Source: Challenges in Scalability and Maintenance

While many healthcare organizations could initially gravitate toward open-source tools under the impression that they are cost-effective, the reality often tells a different story. The high maintenance costs associated with open-source solutions can quickly become burdensome. Challenges such as complex setups, ongoing maintenance requirements, and scalability issues frequently lead our customers to hidden expenses that outweigh initial savings.

Furthermore, as data volumes grow, scaling open-source solutions to meet performance demands can become increasingly difficult, often requiring expensive infrastructure upgrades. 

Both Innovaccer and Tata 1mg have demonstrated that Kloudfuse provides a centralized observability platform specifically designed to optimize costs. With an integrated approach, Kloudfuse consolidates and replaces multiple disparate open-source tools. 

Another important factor is that while Kloudfuse is a hosted platform, similar to open-source options, its Control Plane streamlines deployment and maintenance, effectively making it a self-hosted solution. This provides all the necessary administration and management tools for enhanced reliability, scalability, and ease of use.

Additionally, many open-source tools lack the integrated data lakes required to manage the complexity and volumes of observable data effectively. Kloudfuse’s scalable data lake can handle high volumes and diverse operational data without compromising performance. Graphite Health’s deployment underscores this need; their switch to Kloudfuse enabled them to scale effectively, achieving a fourfold increase in deployment speed while adapting to evolving requirements. 

Conclusion

The experiences of Innovaccer, Tata 1mg, Graphite Health, GE Healthcare, and others in deploying Kloudfuse illustrate the transformative potential of a unified observability approach in healthcare. By learning from these lessons, we hope that organizations can improve on how they architect their observability solutions. 

As the healthcare landscape continues to evolve, we believe that investing in comprehensive, scalable, and secure observability solutions will better position healthtech and healthcare for improved operational efficiency and enhanced patient outcomes.

HIMSS25 next month