September 13, 2023
Yann Leger
@yann_eu
Alisdair Broshar
@AlisdairBroshar
At Koyeb, we're working to build the most seamless way to deploy apps to production without worrying about infrastructure. But there's still plenty to keep you busy at the application layer with performance tuning and troubleshooting. That's why we're introducing Metrics — an easy way to monitor and troubleshoot application performance.
Deploying on Koyeb makes thinking about infrastructure or orchestration unnecessary. We built a serverless engine and a global networking layer so that you can use that time to focus on your apps. Koyeb Metrics is designed to help you understand what is happening with your services at the application level.
Monitoring the resources your applications consume and the requests they receive is key when investigating performance and stability issues. We track CPU and RAM usage as well as inbound request volume, latency, and transfers.
All your applications now benefit from built-in observability of core resources at no additional cost. Metrics are enabled by default and don't require any configuration. The feature is integrated directly into the control panel and accessible for all your services.
When you deploy a service on Koyeb, you can view its metrics as soon as the service instances are running. Metrics are divided into two categories:
Here are the details of the metrics we collect and expose at the instance level:
At the service level, we collect three types of metrics at our ingress gateways. These gateways are responsible for load-balancing the traffic inside of each location. We aggregate the data from all of the ingress gateways that process traffic for your service. If you're curious about what our ingress gateway is doing, we previously wrote about how we build our multi-region service mesh.
The metrics we expose at the service level are:
All of these metrics are consolidated into one dashboard, located in the Metrics tab of the service view.
You can view data for periods of 5 minutes, 15 minutes, 1 hour, 6 hours, 1 day, 2 days, and 7 days with a granularity to the minute during the first hour.
Two important notes:
Having a high-level understanding of what is happening in your services is critical, especially when you need to investigate and diagnose performance issues. Metrics are designed to answer common questions you might face:
These insights help you understand your application better and find the root cause of problems faster. You might want a more advanced observability stack for mission-critical applications, but this visibility on essential metrics should help you troubleshoot common problems.
Metrics is a built-in feature of the Koyeb platform and is now available to all users. Behind the scenes, this feature is powered by Grafana Mimir. We have a blog post coming soon diving into the internals and explaining why we selected Mimir.
We already have some improvements in mind including additional metrics (disk usage or internal data transfer) and a prometheus compatible endpoint to scrape all your metrics. If you would like to see these features, vote for them on our feedback platform.
If this is your first time hearing about Koyeb, welcome to the fastest way to deploy applications globally! Koyeb is a developer-friendly serverless platform. No more DevOps, servers or infrastructure management. We take care of that for you, so you can focus on developing what is important to you.
Koyeb is a developer-friendly serverless platform to deploy any apps globally.
Start for freeDeploy 2 services for free and enjoy our predictable pricing as you grow
Get up and running in 5 minutes