career

Chronosphere.io The Next Adventure

Metrics are the foundation for monitoring the health and performance of the applications that run modern technology solutions today. Site Reliability Engineers (SREs) have used metrics to monitor applications as they have moved from running on servers to running in containers. With the rapid shift of applications run in Kubernetes (K8) containers finding ways to monitor thousands of containers is easy with Prometheus. Prometheus is an open-source metric standard that emits valuable metrics.

Why Change?

Many people have reached out and asked why after an incredible 7+ years at New Relic, I have left to join Chronosphere. The short answer is I had to challenge myself as the industry evolved. Working with some fantastic companies searching for understanding the health of their critical applications, the need for developing monitoring solutions was essential. As applications moved into containers in a concise 4-5 years, the custom Application Performance Monitoring (APM) solutions that worked so well in the past became heavy and cost-prohibitive, running APM agents on thousands of containers the constantly start/stop became unreasonable.

Monitoring as code became a priority for the business and developers alike - incorporating Prometheus metrics into servers, applications, K8, etc.

With this shift in how companies monitor and rapid growth of K8 containers making sense of the metric explosion is where the traditional monitoring and observability companies have been slow to adapt.

Next Generation

The future is evident as we come out of the pandemic technology is not going to slow down. It is just going to speed up. The need to collect and store relevant metrics for applications:

  • Autonomous driving vehicles
  • Renewable energy applications
  • Home automation
  • Advertising
  • Marketing
  • All technology.

Chronosphere was the only company I talked with a proven platform with M3DB that can scale to handle millions of metrics.

Years in the monitoring space have taught me nothing is free. You either have to find ways to monitor less through ratio-based monitoring (for example - I watch one out of every ten servers to save on cost), compression, aggregation, and reduction client-side (increase in CPU and memory). Or you discard the metrics you don’t want to keep and roll up the metrics you want to trend over the long term.

Onward

I am looking forward to the next 3-5 years in technology, making a prediction for myself to move into the renewable energy space, leveraging technology to make the world just a little better. Funny how everything comes back to mathematics, helping people understand through metrics every day!

Subscribe to Tales From The Tech

Get the latest posts delivered right to your inbox