While the 6h in the Chaos DC is valid the 6h in the Light DC is not
The model fetches the highest timestamp (in this case Chaos') and tries to check if it's valid ignoring that Light had an invalid timestamp at the time of the DB fetch
Setting the query to get the min timestamp would cause the Europe results to get zeroed out
Fetching by world and then summing would be too costly as you'd also have to filter per world
This issue is happening on a datacenter scale as well
--------------------EUROPE-------------------------
{"log_timestamp":"2024-01-12 15:07:08","message":"Gilflux ranking for item 16564 on world is OK on range of 7d."}
{"log_timestamp":"2024-01-12 15:07:08","message":"Gilflux ranking for item 16564 on world is OK on range of 3d."}
{"log_timestamp":"2024-01-12 15:07:08","message":"Gilflux ranking for item 16564 on world is OK on range of 1d."}
{"log_timestamp":"2024-01-12 15:07:08","message":"Gilflux ranking for item 16564 on world is OK on range of 12h."}
{"log_timestamp":"2024-01-12 15:07:08","message":"Gilflux ranking for item 16564 on world is OK on range of 6h."}
{"log_timestamp":"2024-01-12 15:07:08","message":"[13065000 > 1705061228000] Gilflux ranking for item 16564 is outdated on range of 3h."}
{"log_timestamp":"2024-01-12 15:07:08","message":"[13065000 > 1705068428000] Gilflux ranking for item 16564 is outdated on range of 1h."}
--------------------LIGHT-------------------------
{"log_timestamp":"2024-01-12 15:07:30","message":"Gilflux ranking for item 16564 on world is OK on range of 7d."}
{"log_timestamp":"2024-01-12 15:07:30","message":"Gilflux ranking for item 16564 on world is OK on range of 3d."}
{"log_timestamp":"2024-01-12 15:07:30","message":"Gilflux ranking for item 16564 on world is OK on range of 1d."}
{"log_timestamp":"2024-01-12 15:07:30","message":"[48684000 > 1705028850000] Gilflux ranking for item 16564 is outdated on range of 12h."}
{"log_timestamp":"2024-01-12 15:07:30","message":"[48684000 > 1705050450000] Gilflux ranking for item 16564 is outdated on range of 6h."}
{"log_timestamp":"2024-01-12 15:07:30","message":"[48684000 > 1705061250000] Gilflux ranking for item 16564 is outdated on range of 3h."}
{"log_timestamp":"2024-01-12 15:07:30","message":"[48684000 > 1705068450000] Gilflux ranking for item 16564 is outdated on range of 1h."}
--------------------CHAOS-------------------------
{"log_timestamp":"2024-01-12 15:15:25","message":"Gilflux ranking for item 16564 on world is OK on range of 7d."}
{"log_timestamp":"2024-01-12 15:15:25","message":"Gilflux ranking for item 16564 on world is OK on range of 3d."}
{"log_timestamp":"2024-01-12 15:15:25","message":"Gilflux ranking for item 16564 on world is OK on range of 1d."}
{"log_timestamp":"2024-01-12 15:15:25","message":"Gilflux ranking for item 16564 on world is OK on range of 12h."}
{"log_timestamp":"2024-01-12 15:15:25","message":"Gilflux ranking for item 16564 on world is OK on range of 6h."}
{"log_timestamp":"2024-01-12 15:15:25","message":"[13562000 > 1705061725000] Gilflux ranking for item 16564 is outdated on range of 3h."}
{"log_timestamp":"2024-01-12 15:15:25","message":"[13562000 > 1705068925000] Gilflux ranking for item 16564 is outdated on range of 1h."}
While the 6h in the Chaos DC is valid the 6h in the Light DC is not
The model fetches the highest timestamp (in this case Chaos') and tries to check if it's valid ignoring that Light had an invalid timestamp at the time of the DB fetch
Setting the query to get the min timestamp would cause the Europe results to get zeroed out
Fetching by world and then summing would be too costly as you'd also have to filter per world
This issue is happening on a datacenter scale as well