Boosting Issue Resolution and User Engagement for a Healthcare Startup with AWS Observability Best Practices

Apollo is a technology startup tackling a two-sided problem in healthcare: unusable data and overwhelming paperwork. Despite widespread adoption of electronic health records, most practices are forced to operate blindly and slowly because their data are disorganized, stale, and incomplete. On the other hand, most clinicians are now on the brink of burnout from spending more time on paperwork than with patients. These problems are bad across healthcare, but they have become especially vicious in cancer care where the complexity and cost of care is highest. The result for patients is delayed care, unnecessary insurance denials, and lack of access to the latest treatments.

Apollo's mission is to enable clinicians to spend more time with patients and less time on paperwork. Apollo's core product is an always-on AI clinical data analyst that reviews and organizes medical records 100x faster than a human. By adding Apollo to their care teams, clinicians can offload repetitive administrative tasks and practices can level up their operations with clean, complete, and real-time data.

The startup sought to establish a comprehensive observability framework that would enable real-time monitoring and rapid issue resolution while maintaining strict compliance standards and protecting Personally Identifiable Information (PII).

As an early-stage healthcare tech startup focused on delivering a seamless clinical experience, Apollo encountered several critical challenges as they began to scale:

Their logging system generated data but lacked organization and readability, making it difficult to diagnose platform issues efficiently. 

Being a high-touch organization, they aimed to personally engage with clinical users, particularly new users in their first few days of using the platform, but tracking detailed user activity in real-time had become cumbersome as the volume of logs grew.

Additionally, with a growing user base, they lacked robust alerting to proactively monitor platform performance. There was limited visibility into key metrics such as latency and downtime leaving them in a reactive state when issues occurred.

Compounding these challenges was their small team size and tight timeline for launch, which left them needing additional expertise to accelerate preparations for their go-live.

To address Apollo’s challenges, we implemented a comprehensive observability solution using AWS CloudWatch. This began with setting up CloudWatch Alerts to monitor critical events in real time, such as application downtime, 4xx/5xx errors, new user logins, and application errors. This gave the team the ability to quickly identify and respond to platform issues and reach out to new users upon their first sign-in.

Next, we developed CloudWatch Dashboards that offered visibility into key operational metrics. These dashboards allowed the Apollo team to monitor EC2 health (CPU and memory utilization), application performance (HTTP request counts and latency), error rates, Beanstalk and environment health, and autoscaling activity. This provided a clear view of system health and allowed the startup to proactively address potential issues before they impacted the user experience.

Recognizing the importance of strict data security in healthcare, we also streamlined access administration and control within AWS Control Tower, ensuring that the solution adhered to strict compliance standards while maintaining the necessary visibility for operational teams.

Lastly, we worked with the startup to validate their AWS architecture, ensuring it was optimized for performance and scalability. This positioned them for long-term success as they continued to grow, giving them confidence in their cloud infrastructure and its ability to handle increased user demands in the future.