Kinsta charges for Application Hosting based on bandwidth, build time, and application hosting pod usage. With Application Analytics, you can get insights into your application’s usage data. You can view analytics for each application in application-level analytics or for all of your applications in company-level analytics.

When viewing analytics, you can choose to see data for the past 24 hours, 7 days, 30 days, or the current billing cycle (Current month in the dropdown menu).

If you highlight a specific point in time on a graph, it will show you some comparison data, such as the percentage difference between each day.

Application Level Analytics

Application-level analytics shows you the following data for your application:

  • Data usage
  • Build time
  • Runtime
  • CPU usage for processes
  • Memory usage for processes
  • HTTP Requests

To access these in-depth reports, go to your application’s Analytics page (Applications > appname > Analytics).

Application-level Analytics in MyKinsta.
Application-level Analytics in MyKinsta.

 

Select each tab under Analytics to see the associated data.

Resources

Under the Resources section, you can view data about your application’s bandwidth, build time, and runtime.

Bandwidth

The bandwidth chart shows the total egress data your application has transmitted. Note: We only charge for egress (traffic from the application to the user); we do not charge for ingress (traffic coming into your application).

Bandwidth chart in application-level analytics.
Bandwidth chart in application-level analytics.

Build time

The build time chart shows you the total time spent building the application.

Build time chart in application-level analytics.
Build time chart in application-level analytics.

Runtime

In the runtime report, you can see the total time the application has been running after being built and deployed, with each process and its associated pod size in a separate row.

Runtime report in application-level analytics.
Runtime report in application-level analytics.

Processes

Under the Processes section, you can view charts for CPU usage and memory usage for each of your application’s processes.

CPU usage

In the CPU usage chart, you can see the average of the total CPU utilization for the selected time period, expressed as a percentage of the instance’s CPU resources. If you see a high percentage of CPU usage (near 100%), we recommend upgrading to an application hosting pod with more CPU for that process.

CPU usage chart in application-level analytics.
CPU usage chart in application-level analytics.

Memory usage

The memory usage chart shows the average of the total memory (RAM) used for the selected time period. If a process uses most or all of the memory available, we recommend upgrading to an application hosting pod with more available memory.

Memory usage chart in application-level analytics.
Memory usage chart in application-level analytics.

HTTP Requests

Under the HTTP Requests section, you can view charts for HTTP requests, HTTP P95 latency,  and HTTP P50 latency.

HTTP requests

The HTTP Requests chart shows the number of HTTP requests for the time period selected and the average requests per minute (RPM).

HTTP Requests chart in application-level analytics.
HTTP Requests chart in application-level analytics.

HTTP P95 latency

HTTP P95 latency is the value at which 5% of the requests exceed the duration of (are slower than) the P95 latency threshold of all requests for the duration selected.

For example, if the 95th percentile threshold (response time) is 907.8 ms and there were 100 requests during the time period selected, 95 out of 100 requests are 907.8 ms or faster, and 5 out of 100 requests are slower than 907.8 ms.

The HTTP P95 latency chart shows the average P95 threshold for the duration selected under the chart name. Each point on the chart shows the P95 threshold for that specific hour or day, depending on the duration selected.

HTTP P95 latency chart in application-level analytics.
HTTP P95 latency chart in application-level analytics.

HTTP P50 latency

HTTP P50 latency is the value at which 50% of the requests are slower, and 50% are faster than the P50 latency threshold of all requests for the duration selected.

For example, if the 50th percentile threshold (response time) is 802.2 ms, half of the requests are 802.2 ms or faster, and the other half are slower than 802.2 ms.

The HTTP P50 latency chart shows the average P50 threshold for the duration selected under the chart name. Each point on the chart shows the P50 threshold for that specific hour or day, depending on the duration selected.

HTTP P50 latency chart in application-level analytics.
HTTP P50 latency chart in application-level analytics.

Company Level Analytics

Company-level analytics shows you data usage, build time, and runtime for all of your applications. To access these reports, go to your company’s Analytics page (Your name > Company settings > Analytics). Note: If you also have Databse Hosting or WordPress Hosting, you’ll need to select the Applications subpage to view your company-level analytics.

Bandwidth

The bandwidth chart shows the total egress data your applications have transmitted for the selected time period. Each color block represents an individual application or group of applications, so you can see which application uses the most or the least amount of bandwidth. If you have 4 or more applications, the applications that generated the lowest bandwidth are grouped as “Others.”

Bandwidth chart in company-level application analytics.
Bandwidth chart in company-level application analytics.

Build time

The build time chart shows you the total time spent building all of your applications, grouped by build machine type.

Build time chart in company-level application analytics.
Build time chart in company-level application analytics.

Runtime

In the runtime report, you can see the total usage time all of the applications run after being built and deployed, with each resource type and count in a separate row.

Build time chart in company-level application analytics.
Build time chart in company-level application analytics.

Additional Notes

Analytics data is retained for 30 days. We suggest checking your Application Analytics frequently after first deploying to Kinsta and periodically after that. If you see any unexplained analytics data or inconsistency that concerns you, let our Support team know, and we can further investigate to help determine the cause.