Small gaps in geo_country enrichments

For engineers Enrichment
1

I observe a picture where for the same user with the same user_ipaddress the geo fields are not enriched. For example, for a few events (and minutes) everything was fine, then the event (or events) was not enriched (nulls in geo attributs), then everything was fine again.

I’ve noticed this for just a few (5-10) user_ipaddress per day, but it bothers me that I can’t explain why this might be happening.

For enrichment geo_country we use GeoIP2 City Database

2

I haven’t really seen this happen before - and this seems less likely to happen particularly as the LRU cache should be storing these (unless it’s a very high volume pipeline) for quite some time.

Are you running multiple versions of enrich, what configuration do you currently have setup for the GeoIP enrichment?

3

Hello, Mike!

We use group of enrichers with the same version of enrich.

Configuration:

{
	"schema": "iglu:com.snowplowanalytics.snowplow/ip_lookups/jsonschema/2-0-0",

	"data": {

		"name": "ip_lookups",
		"vendor": "com.snowplowanalytics.snowplow",
		"enabled": true,
		"parameters": {
			"geo": {
				"database": "GeoLite2-City.mmdb",
				"uri": "s3://our_bucket/maxmind_new"
			}
		}
	}
}
4

Do you use the assetsUpdatePeriod setting in your enricher config? I’m not aware of any issues with that feature, but if there’s a cadence it might line up?

5

jethron, yes, assetsUpdatePeriod is one day

6

We think it could happen because of random start time of rise machine in group: one machine is updated and can enrich event, another new machine rises, does not have time to update the ip base and does not enrich.