Getting started

Integration & migration

Image & video API

Video Player SDK

DAM user guide

API overview

Account

Usage Analytics

Learn how to access and understand usage analytics in ImageKit.io

•

ImageKit provides detailed usage analytics to help you understand how your account is being used. To access usage analytics, log in to your ImageKit dashboard and navigate to the Usage Analytics section.

Note
If you use sub-accounts, the parent account can only show an overview of analytics, while detailed analytics are available in the sub-accounts.

The overall usage summary looks like this.

Let's explore different sections of usage analytics.

Daily usage graphs for bandwidth, requests, VPUs, and extensions

Monitor bandwidth, requests, VPUs, and extensions with daily usage graphs. These are usually updated within a few hours of actual consumption.

Within VPU, you can also see the split for different resolutions, such as SD, HD, 4K, etc.

VPU analytics

Go to VPU analytics to see what consumed your video processing units (VPUs), broken down by day, hour, and asset, down to every individual job.

It covers every job that consumes VPUs: video transformations, GIF to video conversions, adaptive bitrate streaming (ABS), and thumbnail generation. Usage usually appears within a few minutes of processing. All dates and times are in UTC.

Summary

  • VPUs consumed: The total VPUs consumed in the selected period.
  • Billed duration: The video time your VPUs are charged on. For a transformation, this is the length of the output video, counted once for every smart crop requested. An ABS manifest and a thumbnail each cost a flat 30 seconds, since neither depends on the length of the video.
  • Video output produced: The total length of video actually produced, measured on the output. A trimmed video counts its trimmed length, not the length of the original. ABS manifests and thumbnails are not counted, as they cost a flat fee.
  • Jobs processed: The number of jobs that consumed VPUs.

The VPUs consumed by each job type are shown below the summary.

Daily VPU consumption

A chart and a table of VPUs consumed each day, split by job type. Select a day to see its usage by hour and the assets processed on that day.

VPU by codec and resolution

Shows billed duration × multiplier = VPU for each codec and resolution. The rows add up to the total for the period. AV1 costs 10 units per second compared to 1 for H.264 and VP9, so the codec you deliver in usually explains a jump in usage.

Adaptive bitrate ladders

Lists each ABS stream with its asset URL, master playlist URL, number of renditions, and VPUs. Select a stream to see the resolution, dimensions, codec, and duration of each rendition generated for it. The total includes the flat 30 seconds charged for the master playlist.

A rendition that was already generated by an earlier request is served from the cache. It is not processed or charged again, so it does not appear under the new stream.

Assets consuming VPUs

Lists the videos that consumed the most VPUs. Search by URL to find a specific video (type at least 3 characters). Select a video to see every job run on it, with the output URL, resolution, codec, format, streaming protocol, output duration, billed duration, multiplier, and VPUs for each job.

How VPU figures add up
When billing, ImageKit rounds billed seconds up to the next whole second once for each codec and resolution over the whole billing period. VPUs consumed and the VPU by codec and resolution table follow the same rule over the selected period, so for a full billing period they match your invoice.

Figures for a single day, hour, or asset are rounded up on their own. A rounded figure cannot be split back across assets, so these figures can add up to a few units more than the total. The VPU by codec and resolution table has the exact numbers.

Extension analytics

Go to Extension analytics to see what consumed your extension units, broken down by day, hour, and asset, down to every individual operation.

It covers every operation that consumes extension units, such as background removal, AI tagging, upscaling, and Creative Automation variations. Usage usually appears within a few minutes of processing. All dates and times are in UTC.

Summary

  • Extension units consumed: The total extension units consumed in the selected period. This is your billed figure.
  • Operations: The number of times an extension ran. Units per operation differ between extensions, so more operations does not always mean more units.
  • Highest consuming: The extension that used the most units in the period.
  • Extensions used: The number of different extensions used.

Daily extension units

A chart and a table of extension units consumed each day, split by extension. Select a day to see its usage by hour and the assets involved.

Units by extension

Shows operations × units per operation = extension units for each extension. The rows add up to the total for the period. Units per operation differ a lot between extensions, so this table usually explains a jump in usage.

Assets consuming extension units

Lists the assets that used the most extension units, by URL or by file in the Media Library. Search by URL, file path, file ID, or extension name (type at least 3 characters). Select an asset to see every operation run on it, with the extension, request, units per operation, and extension units for each operation.

Note
Totals are calculated at your current rate for each extension, which is how you are billed. If an operation was recorded at a different rate, it is marked in the operation list.

Top assets by bandwidth and requests

Under Media use analysis, you can see the top assets by bandwidth and requests. We provide split for images, videos, and other assets based on the Content-Type response header.

This lets you determine which assets consume the most bandwidth and are requested frequently. For example, if you find a few large-sized GIFs consuming a lot of bandwidth, consider converting them to videos to save bandwidth. Or, for example, if you find a lot of high-resolution images, consider resizing them correctly.

Error reports

Go to the error reports section to see the errors received on your ImageKit URLs in the selected duration.

If you see a lot of 404 errors, look at the top URLs and the reasons to fix them in your application. Learn more about troubleshooting 404 errors.

Top image and video transformations

This section summarizes the top transformations applied to images and videos. Ideally, for high performance, a few transformations should account for most of the request volume.

If you see too many transformations here, each with a very low percentage usage and a low cache hit ratio on the CDN, consider consolidating your transformations to a smaller number.

If you suspect a transformation is incorrect or should not have been used, first check your code for it. If still an issue, consider using named transformation and restricting un-named transformations for more control on the transformations that can be applied to your media assets.

Cache hit statistics

This section provides information about cache hit statistics. It shows the number of cache hits, misses, and error requests on the CDN integrated in ImageKit.

A cache hit percentage above 90% is considered excellent, while above 80% is good. A lower cache hit ratio means the cache is not utilized effectively. This is usually the case if you are using a lot of unique URL transformations, signed URLs with expiry, or the cache is being bypassed due to random query parameters containing a timestamp or a unique identifier.

Top usage by countries, devices, browsers, and referrers

This section helps you identify the top usage by countries, browsers, output formats, devices, user agents, referrers, and IP addresses. Looking at anomalies here can help identify, optimize, and fix high usage on ImageKit.

  • By countries: determined from the country of CDN edge used to serve content and may not match the country of your user
  • By browsers: Determined from the User-Agent string value in the request header.
  • By formats: Determined from the value of the HTTP Content-Type header of the response.
  • By devices: Determined from the value of the User-Agent string in the request header
  • By user agents: Determined from the value of the User-Agent header in the request header.
  • By referer: Determined from the Referer header in the request headers.
  • By IP addresses: Determined by the IP address of the incoming request.

If you observe hotlinking or unauthorized usage, you can use the advanced security features to block specific IPs, referrers, or user agents.

Suppose you observe significantly high usage for a single format like JPEG or PNG. In that case, the images are not optimized to a modern format, like WebP or AVIF. Consider checking the image and video optimizations and ensuring they work across all websites and apps.