Ultimately this is a more semantic format and allows for better parsing of data. Logs should probably be saved as JSON (still as a rolling file), and then ingested periodically into ElasticSearch, and analyzed with Kibana. The files could be synced with SyncThing and just ingested before analysis. This is needed instead of having a constantly running ElasticSearch/Kibana instance, because those actually take a decent amount of resources to run (~4GB for ElasticSearch?).
Possibly also upload a copy of logs to Seq? Probably not appropriate for more "formal" bots (e.g. Polybius, Puck), but having a backup-ish copy always accessible on Seq would be convenient. (this might need its own issue)
(doesn't necessarily depend on #390, but it'd be nice to do it first to avoid repeating work)
Ultimately this is a more semantic format and allows for better parsing of data. Logs should probably be saved as JSON (still as a rolling file), and then ingested periodically into ElasticSearch, and analyzed with Kibana. The files could be synced with SyncThing and just ingested before analysis. This is needed instead of having a constantly running ElasticSearch/Kibana instance, because those actually take a decent amount of resources to run (~4GB for ElasticSearch?).
Possibly also upload a copy of logs to Seq? Probably not appropriate for more "formal" bots (e.g. Polybius, Puck), but having a backup-ish copy always accessible on Seq would be convenient. (this might need its own issue)
(doesn't necessarily depend on #390, but it'd be nice to do it first to avoid repeating work)