How integration works
EINBLIQ.IO analyses playback from your CDN logs, from CMCD v2 events sent by your players, or from both combined.
Updated
EINBLIQ.IO supports two integration methods. Log analysis uses the access logs of your CDN. Event analysis uses CMCD v2 events that your players send to the EINBLIQ.IO CMCD v2 endpoint. Each method works on its own; combining both gives the most complete data.
| Log analysis | Event analysis | |
|---|---|---|
| Data source | CDN access logs, optionally with CMCD request data | CMCD v2 events reported by the player |
| Sent by | Your CDN, or queried by EINBLIQ.IO from your log analytics platform | Your players |
| Configured in | Log configuration of your CDN | Player configuration |
| Guides | CDN guides | Player setup |
Log analysis through your CDN
CDN logs record segment, manifest and licence requests that pass through the CDN and are included in its logging configuration. SPECTRA groups suitable records into playback sessions, so FOCUS can discover unusual patterns as Signals and investigate each one. This works with suitable CDN logs alone, even without CMCD or an explicit session ID. Playback duration and startup behaviour derived from logs are estimates; direct player errors and stalls need player telemetry.
When your players send CMCD (Common Media Client Data, CTA-5004) in request mode, each request also carries what the player reported at that moment, such as buffer length, requested bitrate and a session ID. The CDN writes these values into its log line next to the delivery data. No proprietary client-side SDK is needed: CMCD is an open standard that common players already implement.
If your CDN logs already flow into a log analytics platform such as Hydrolix or Akamai TrafficPeak, EINBLIQ.IO queries the log table there instead, without log delivery or changes to your CDN. See Hydrolix and Akamai TrafficPeak.
Required log fields
Field names differ between CDNs. What matters is that each log line contains these values:
| Field | Content |
|---|---|
| Timestamp | Request time in UTC, ideally with milliseconds |
| Hostname | Requested host |
| Client IP | IP address of the requesting client |
| Edge IP | IP address or identifier of the edge server |
| User agent | HTTP User-Agent header |
| URI | Path of the requested object |
| Query string | Full query string, including CMCD= when present |
| Status code | HTTP response status |
| Bytes | Response size in bytes |
| Duration | Time to serve the request, in milliseconds |
| Cache status | Hit, miss or the CDN’s equivalent |
| Referer | HTTP Referer header |
| CMCD (optional) | Player-reported values, from the query string or captured CMCD request headers, when available |
Event analysis through the EINBLIQ.IO CMCD v2 endpoint
CMCD v2 adds event mode: the player sends batches of events, such as play state changes, errors and periodic status reports, to a reporting endpoint, independently of segment requests. EINBLIQ.IO provides a CMCD v2 endpoint for this. Its address and any required token are provided during onboarding; you add the endpoint as the event target in your player configuration. Event analysis needs no changes to your CDN.
Event mode is available in dash.js and hls.js, and as an experimental feature in Shaka Player. The player setup page shows where the endpoint goes in each player.
Combining both methods
Combining both methods gives the most complete data. CMCD v2 events report what happened in the player, such as play state changes, rebuffering and errors. CDN logs add the delivery details of every request: status code, cache status, response time, bytes and the edge server that answered. With both, an investigation can follow a playback problem from the player back to the requests behind it, which is often what identifies the root cause.
Either source on its own is enough to start, and the other can be added later.
SPECTRA joins the events with your CDN logs through the session ID (sid) and, where known, the content ID (cid). If you set these IDs yourself, use the same values in request mode and event mode.
Prerequisites
- For log analysis: editing rights for the log configuration of your CDN. You set up the log delivery yourself; EINBLIQ.IO does not need access to your CDN. With Hydrolix or Akamai TrafficPeak, read access to the log table replaces the log delivery.
- For event analysis: access to your player configuration.
- The delivery target, the CMCD v2 endpoint address and their credentials, which EINBLIQ.IO provides. The guides show these values as highlighted placeholders such as
<ACCESS_KEY_ID>.
Integration steps
- You tell EINBLIQ.IO which integration methods you plan to use and, for log analysis, which CDN or log pipeline and which services should be included.
- EINBLIQ.IO prepares the delivery target or endpoint and sends you the values for your guide.
- For log analysis, you configure log delivery as described in the guide for your CDN, or provide read access to your log table.
- In your players, you enable CMCD: request mode adds player context to the logs, ideally as a query parameter, and event mode reports to the EINBLIQ.IO CMCD v2 endpoint.
- EINBLIQ.IO confirms that data arrives with the expected fields and reviews the first sessions with you.
Guides by CDN
- AkamaiDataStream 2 to an S3 bucket, including the dedicated CMCD field.
- Amazon CloudFrontStandard logging (v2) to an S3 bucket in the EINBLIQ.IO AWS account.
- CloudflareLogpush to an S3 bucket, with CMCD headers as custom log fields.
- FastlyAmazon S3 logging endpoint to an S3 bucket, with CMCD from query string and headers.
- GcoreLogs Uploader to an S3 bucket, with optional CMCD header logging.
- KafkaKafka Connect S3 sink for logs you already collect in Kafka.
- Other CDNs and your own cachesThe general requirements for any CDN or cache.
- Hydrolix and Akamai TrafficPeakDirect queries on the log table you already collect, without log delivery.