Clarus: We were notified of the AWS outage in time so that none of our patient calls were affected

• Since switching to serverless and Dashbird, Clarus’ engineering team is spending resources on building the product rather than wasting time on messy and archaic logs.

• Dashbird’s Well-Architected Lens ensures everything is set up and running correctly, proving to be an effective, proactive tool to ensure there are no issues rather than solving problems reactively. 

• Clarus’ API request response time is 123% quicker than their previous infrastructure. Monthly infrastructure costs went down by 301% since switching to serverless. 

• Zero patient calls affected during AWS biggest outage in November 2020.

Starting point

Clarus is a complete call management solution for healthcare providers, doctor’s offices, healthcare groups and hospitals. They have two services that rely on different AWS products: Daytime and After-Hours. 

Clarus replaces the traditional medical answering service and gives medical staff and providers the tools to manage after-hours calls from anywhere. The Clarus platform enables the management of urgent calls to physicians and doctors faster and quicker than other solutions in the market, with privacy features to keep the doctors’ and physicians’ numbers secure. Patients also have to be able to get through to their doctors in a timely manner—they have to be able to call in 24/7, 365 days a year. 

Being in the healthcare industry and in a position where the platform has to relay critical and time-sensitive information to the right people means uptime and monitoring are top priorities. 

Clarus previously used a HIPAA-compliant development company that managed everything. 

This meant that there were three immediate concerns: 

1. Scaling – a rush of incoming calls during peak business hours meant the previous infrastructure could not support the incoming requests.
2. Overall management – a third-party HIPAA-compliant development service meant Clarus’ internal teams did not have access to their own infrastructure.
3. Visibility – having little visibility meant when services went out, the development team had to answer calls at odd hours and had trouble knowing what to fix. Going through logs and finding the actual issues were not only time-consuming but also morale-draining.

After bringing the development in-house along with migrating to serverless and Dashbird, the Clarus team had built an efficient process for DevOps and customer success. Today, they no longer have to wake up at 5 in the morning for service outages, and no longer manually monitor resources since Dashbird instantly alerts the team of any issues.

Serverless is awesome because I don’t have to worry about servers and I can just focus on developing the actual products.

– Erik Galloway, Lead Engineer at Clarus

The challenge

Clarus has three Lambdas in use: 

1. A web app that is the call management platform handling communication from clients to Twilio to their infrastructure. 
2. Queue workers that handle notifications to doctors, which leverage AWS SQS service. 
3. Scheduled task Lambda which runs every minute, managing the notification pipeline, sending queued jobs to SQS.

We needed something that could monitor everything from a distance without having to be so deeply woven into the actual code and the application.

– Erik Galloway, Lead Engineer at Clarus

With various AWS products in use, Clarus needed a serverless observability tool — one that could work out of the box and without additional development into the existing infrastructure. 

Another immediate challenge was analyzing logs efficiently; looking at logs, searching them, and having a quick overview of everything. Since Dashbird offers detailed logs with traces along with the Well-Architected Lens, there is simply no competition between Dashbird and AWS’ default UI. 

The last challenge was being notified of custom events. Having different types of alarms allows Clarus’ team to be notified of specific events that they need to keep an eye on. 

We were able to analyze our data, debug, and get alerted of important events and errors immediately after signing up.

– Erik Galloway, Lead Engineer at Clarus

Dashbird saves the day during AWS biggest outage in 2020

Clarus has two golden rules, especially in engineering:

1. Patients and doctors need to be able to communicate 24/7, 365 days a year. 
2. Doctors must be able to be notified of new calls so they can respond back to the patient as quickly as possible.

For the organization, uptime and working communication are what drives the business. If the two aren’t happening then it’s the biggest emergency that Clarus can have. 

Unfortunately, AWS had a multi-hour outage in November 2020, which affected a significant chunk of the internet, including Clarus.

Thankfully, Clarus had Dashbird running and Erik and the team were notified right away. They were able to get to the issue so quickly that no patient calls were affected by the outage.

Dashbird is really our all encompassing monitoring tool for all services at once.

– Erik Galloway, Lead Engineer at Clarus

Dashbird’s Well-Architected Lens provides confidence

The Well-Architected Insights uses AWS’ Well-Architected Framework to ensure everything is set up correctly and operating smoothly.

It is very reassuring to me to have all of those different insights.

– Erik Galloway, Lead Engineer at Clarus

The Well-Architected Insights lays out mission-critical data in an easy-to-digest format with the five pillars of the AWS Well-Architected Framework in mind:

• Operational excellence
• Security
• Reliability
• Performance efficiency 
• Cost optimization

By having these five pillars as the core foundation, Dashbird can quickly check and see if there are improvements to be made—whether it’s to meet the standard or for increasing customer demand on the production side.

It’s still really nice to know that there’s nothing wrong with our setup at the moment. Or if there is, I will be alerted that, “Hey, you need to tweak this setting or turn this on now”

– Erik Galloway, Lead Engineer at Clarus

The solution

I have the confidence that everything is okay. If something isn’t okay I am able not only to be alerted by Dashbird, but also to see the location of the exact issue.

– Erik Galloway, Lead Engineer at Clarus

Clarus connects patients in need of care with the right healthcare professionals, leveraging AWS Lambda, API Gateway, DynamoDB, and SQS. The volume itself is a challenge, and having a responsive service with no outage is simply a must for the sake of the patients in need of care.

Since joining Dashbird and switching to serverless, Clarus has improved API request response time by 123% and reduced monthly infrastructure by 301%. Erik experienced a significantly less stressful work environment since Dashbird keeps the systems in check, and provides the confidence so the team could focus on building the product.

We literally automated away the most frustrating part of our team’s tasks. That was not having to manage the servers by going serverless and not having to worry about servers going down or serverless things not working properly because of Dashbird.

– Erik Galloway, Lead Engineer at Clarus

About Clarus

Clarus is an innovative technology company changing the way patients interact with doctors and their staff. By providing automated, cloud-based communication tools to physician practices throughout the United States, Clarus elevates the patient experience and streamlines communication. The company’s unique offering eliminates unknowns and errors of traditional answering services. Their predictable pricing and easy-to-use platform allows provider organizations of any size to transform the way they take calls. To learn more about Clarus, visit claruscare.com.

About Erik Galloway

Erik Galloway is the Lead Engineer of Clarus. Having experience in building world-class applications and development teams, he joined Clarus to elevate the platform to the next level. Erik has vast experience designing and developing large-scale platforms using best practices and led Clarus’ migration to serverless.