Closed dajuno closed 8 months ago
Olaf: (für Bibliothekare) Subscription Zugriffszahlen, Wieviele Nutzer/Institutionen sind dabei (nach Ländern), DSGVO abwägen. Johannes (Idee): Heatmaps, weg durch die Dokumente per Nutzer
https://plausible.io/ looks pretty nice actually, especially from a privacy perspective. quote "P.S. We're completely independent, self-funded, bootstrapped and debt-free. We're not interested in raising funds or taking investment. We choose the subscription business model rather than surveillance capitalism." && "No cookies and fully compliant with GDPR, CCPA and PECR. Made and hosted in the EU, powered by European-owned cloud infrastructure" && open source self-hosted version is completely free
as suggested by Tim, a log analyzer might be a more adequate choice, since we have access to the apache logs. hence, installed goaccess.io
I would like to give mamoto a shot. In theory the main problem will be to tell mw the IP of traefik so that it trusts it's XFF IP forward headers. https://m.mediawiki.org/wiki/Manual:$wgCdnServers Maybe we need a persistent IP if the server name does not work. Otherwise it seems straight forward. See https://github.com/matomo-org/docker and https://m.mediawiki.org/wiki/Extension:MatomoAnalytics
wiki metrics such as number of items/pages etc should be tracked for prometheus/grafana, or, alternatively, statsd+graphite.
TODO:
Metrics scraper and visualization:
List of services of interest:
system monitoring:
usage analysis
publishing
external tools
Metrics for our own tools:
see https://grafana.wikimedia.org/ for inspiration very rough statistics can be found here: https://portal.mardi4nfdi.de/wiki/Special:Statistics