The Peer5 dashboard gives you an overview of what is currently going on in your service from Peer5's point of view. The dashboard contains several graphs and metrics, each highlighting a different aspect of the end-user's experience.
Before we delve in to the graphs themselves, its important to know the terminology used in the description of those graphs:
- End-User: A browser session (i.e., a browser tab). A browser refresh creates a new session.
- Activity: HTTP or P2P Download.
- Watching: Playing a video or rebuffering.
Ordered from left to right, top to bottom¶
HTTP Delivered - Total number of bytes that were downloaded by HTTP from the CDN (or origin server) to end-users.
P2P Delivered - Total number of bytes that were downloaded by P2P from other end-users via the Peer5 peer-to-peer network.
P2P Offload - The percentage of bytes that were delivered in P2P, out of the total bytes that were delivered to end-users. Calculation:
(P2P Delivered) / ((P2P Delivered) + (HTTP Delivered)).
Total Delivered - The total number of bytes that were downloaded to end-users, either via HTTP or P2P. Calculation:
P2P Delivered + HTTP Delivered.
HTTP vs P2P Speed - The graph shows 3 series:
- P2P Ratio (Green line) - The "P2P Offload" for the specified time bucket (1m, 5m, 10m, 1h, etc).
- HTTP Gbps (Blue bars) - The average HTTP bandwidth for the specified time bucket.
(HTTP-Delivered-for-time-bucket-in-Gb)/(Number-of-seconds-in-time-bucket). Looking at long time buckets will cause the average speed to appear lower than it was.
- P2P Gbps (Orange bars) - The average P2P bandwidth for the specified time bucket.
(P2P-Delivered-for-time-bucket-in-Gb)/(Number-of-seconds-in-time-bucket). Looking at long time buckets will cause the average speed to appear lower than it was.
Concurrent Viewers - The average number of end-users that had any activity (HTTP Delivered or P2P Delivered) in the specified time bucket. Looking at longer time buckets will cause the average number of end users to appear lower than it was.
Total playback sessions - The total number of end-users that started a video playback session in the selected time range. It's possible that the selected time range will miss some playback sessions because those sessions started prior to the selected time range start time.
Average Bitrate - The average number of Kilobits per second of video loaded by the player across all end-users. Calculation:
Average session duration - The average amount of time an end-user spent watching a video. Calculation:
(total-time-played + total-time-rebuffered)/(total-playback-sessions). It's possible that the selected time range will miss some playback sessions causing this metric to be misleading in that time range.
Rebuffer % - The percentage of time end-users spent rebuffering a video out of the total time they spent watching it. Calculation:
(total-time-rebuffered)/(total-time-rebuffered + total-time-played).
Played vs Rebuffered - The graph shows 3 series:
- Ratio (Blue line) - Same as "Rebuffer %", just measured for the specified time-bucket.
- Played (Green bars) - The total number of minutes that were spent playing a video (actually playing, not rebuffering) by end-users for the specified time bucket.
- Rebuffered (Red bars) - The total number of minutes that were spent rebuffering a video by end-users for the specified time bucket.
Top 10 active pages (Concurrent viewers) - Same as "Concurrent Viewers", but measured over the top 10 pages for the selected time range. The top 10 pages are chosen by the total number of “concurrent viewers” for those pages during the selected time range.
Peer5 only works on browsers that support WebRTC Data Channels. On browsers that don’t have such support, Peer5 will be automatically disabled and fallback to native playback (Just the same as without Peer5). Those unsupported-browsers will not have any analytics collected from them (Such as HTTP/P2P/UX metrics), which can cause some discrepancy between the numbers that you see on your own analytics systems and the Peer5 one. This discrepancy is expected and will vary based on the % of unsupported-browsers your end-users use.