Event fingerprint enrichment with custom context


#1

Hi,

We want to use the event fingerprint enrichment to deduplicate events.

The documentation says

The input values for this enrichment are taken from all the querystring

Does it mean this enrichment takes into account custom contexts ?
If yes is it possible to exclude some fields of a custom context when computing the event fingerprint ?

Thanks,
Arthur


#2

The documentation for this enrichment is here:

By default, all custom contexts are considered as part of the fingerprint.

Is there a reason why you would want to treat certain contexts as not part of the distinctiveness of the event? This would mean that you would want to treat events with different contexts as duplicates of each other? Having trouble imagining the use case.


#3

Hi @alex,

We have a specific use case where the devices generating the events don’t send the events directly to the snowplow collector.
Instead they send zip files containing, among other things, the events we want to track with Snowplow. Those zip files are stored on S3 and we were thinking of adding a reference to the archive an event originates from in a custom context.

We can do without it though so we’ll just remove this property from the custom context so we can use the event fingerprint enrichment.

Thanks a lot for your reply,
Arthur


#4

Hi @abrenaut - ah, that makes sense, thanks for clarifying.

Yes, unfortunately at the moment you wouldn’t be able to exclude that property from the uniqueness. However, it’s a nice idea. Added:

https://github.com/snowplow/snowplow/issues/3692