Log IPv4/ location data and serve it in the public dashboard while preserving privacy

[PatrickNercessian]

Sub-issue to https://github.com/filecoin-station/roadmap/issues/97

[PatrickNercessian]

We previously discussed that CloudFlare has optional location headers that we can include in requests to spark-api.

https://developers.cloudflare.com/rules/transform/managed-transforms/reference/#add-visitor-location-headers

Thus, in spark-api, we can store this data in InfluxDB, to be consumed by spark-observer. This is similar to the approach discussed in #98.

The end goal is to have a public visualization of geolocation of nodes (while preserving privacy), and a dashboard of the total number of IPv4 subnets over time.

[PatrickNercessian]

There are multiple map visualization tools out there. Grafana has a mapping visualization function . Another common one seems to be Leaflet.js. There are also multiple python visualization tools, and some proprietary services we can use as well, such as Tableau or Google.

In any case, most of these do not seem to have distinct functionality to use ISO-2 region codes, such as the ones given by cloudflare's cr-region-code header. Thus, it seems helpful to store all relevant information to avoid cases where we change our mind on the visualization tool. See https://github.com/filecoin-station/spark-api/pull/339

[PatrickNercessian]

Any thoughts on the above visualization tools? @bajtos @juliangruber @patrickwoodhead

[juliangruber]

While we use Grafana, let's just use what Grafana offers?

[bajtos]

Closing as done.

https://grafana.filstation.app/public-dashboards/2751d44f96914bb3a030b1de13224d19?orgId=1