Infrastructure metrics
Back to home
On this page
Upsun projects are accompanied by live infrastructure metrics that provide an overview of resource usage for environments.
Within the Console, metrics can be found for an environment under Resources.
The information under Resources shows usage metrics for:
Upsun environments: your service, app, and worker containers. These metrics are available for all of your environments.
Default thresholds
All of the graphs show color-coded lines for the following thresholds:
- Usage up to 100% results.
- Usage that crosses 80% results.
- Usage that crosses 90% results.
- Usage that crosses 50% results.
Recommendations
The default thresholds aim to give you an idea of when your hosts or containers are close to running out of resources. The impact differs based on your specific apps and service.
If the resources are hovering close to the 100% threshold, you might want to consider:
- Optimizing your code (if possible)
- Changing your app size or service size
Time intervals
Measurements are taken for each metric every 1 minute. You can select a time frame over which to see these measurements for the entire Resources view. In the primary three views, averages are shown over larger intervals.
View | Time between measurements | Example |
---|---|---|
The last 15 minutes (15m) | 1 minute | 10:00, 10:01, 10:02 |
The last hour (1hr) | 1 minute | 10:00, 10:01, 10:02 |
The last 8 hours (8hr) | 10 minutes | 10:00, 10:10, 10:20, 11:00 |
To zoom in on smaller intervals, select specific ranges in a graph.
The interval between measurements then changes based on the range you choose.
View | Time between measurements |
---|---|
< 2 hours | 1 minute |
2 to 5 hours | 5 minutes |
5+ hours | 20 minutes |
Longer time intervals
You can access historical data for up to 30 days. These data should help you understand trends over time and whether a given measurement is normal, something that occurs occasionally, or a true anomaly.
To see data over a given time frame, use the date picker to select the range to display.
Pressure metrics
CPU pressure is available in your metrics dashboard alongside existing resource metrics such as CPU, memory, and disk usage. This metric helps you spot hidden performance bottlenecks, especially in shared environments where multiple containers compete for resources.
What is Pressure?
Pressure indicates the percentage of time your app’s tasks were delayed due to resource contention. This means they needed CPU, memory, or disk access but had to wait because those resources were fully occupied. For example, 6.8% pressure means that tasks were blocked from making progress for 6.8% of that period of time.
How to understand Pressure
High pressure indicates performance bottlenecks, which means that your app’s tasks are frequently waiting for access to resources, even if overall usage appears normal.
Low pressure with high resource usage typically means your app is making efficient use of its allocated resources. It’s fully utilizing them without significant delays.
Understanding the values
- 0% pressure: No contention - tasks had immediate access to resources.
- 0–100% pressure: During each second of measurement, the application experienced contention for that percentage of time.
- >100% pressure: The app is consistently demanding more resources than are available.
For example, 200% CPU pressure with one CPU already at 100% utilization suggests the app would need roughly three CPUs in total to eliminate contention.
Pressure below 100%
Pressure below 100%, especially when it fluctuates, does not cause noticeable performance issues.
Deployments
Sometimes deployment activities (pushes, merges, syncs) can influence the underlying infrastructure. You don’t want to confuse a spike caused by a successful deploy with an issue that needs your attention.
To see how deployment activity influences the infrastructure, turn on Deployments on your metrics.
Each deployment activity appears as a line on each graph corresponding to the time the activity finished.
- To see information about the deployment activity, hover over the line.
- To see the build log for a specific deployment activity, click the line.