`bungeecord_online_players` gauge metrics accumulate to unusually high values

[weihao]

`bungeecord_online_players` gauge metrics accumulate to unusually high values

[weihao]

i just published a minimal and self-contained setup: https://github.com/weihao/docker-minecraft-network

and I can't reproduce the error, let me know if you catch something

[weihao]

Could you confirm you are using the latest bungeecord / waterfall? what version is it? I can't seem to reproduce it.

[kory33]

Could you confirm you are using the latest bungeecord / waterfall? what version is it? I can't seem to reproduce it.

We are using itzg/bungeecord:java17-2022.1.0 for the BungeeCord deployment with TYPE=WATERFALL, and the container log says that the waterfall version is 1.18-488, which is the current latest.

2022-06-06 04:33:58	
[init] Resolving type given WATERFALL
2022-06-06 04:34:12	
[init] Downloading https://papermc.io/api/v2/projects/waterfall/versions/1.18/builds/488/downloads/waterfall-1.18-488.jar
2022-06-06 04:34:14	
[mc-image-helper] 19:34:14.759 INFO  : Downloaded /server/waterfall-1.18-488.jar

We publish the whole setup (base and overlay), but if you still can't reproduce the issue, I could try minimizing to a minimal self-contained setup.

[weihao]

investigating.

one thing I noticed is why did you turn off ping_passthrough?
https://github.com/GiganticMinecraft/seichi_infra/blob/7a89624763eab97788abeff83dda2dc7a8a9b849/seichi-onp-k8s/manifests/seichi-kubernetes/apps/seichi-gateway/bungeecord/bungeecord-config-patch.yaml#L76

[weihao]

i can't get your setup up.

https://github.com/GiganticMinecraft/seichi_infra/blob/7a89624763eab97788abeff83dda2dc7a8a9b849/seichi-onp-k8s/manifests/seichi-kubernetes/apps/seichi-gateway/bungeecord/bungeecord-plugin-configs-patch.yaml#L31-L32

it can't find redis deployment

[inductor]

@weihao currently we still use on-prem vm redis and we are planning to migrate it to containers.
Ref. GiganticMinecraft/seichi_infra#462

[kory33]

investigating.

one thing I noticed is why did you turn off ping_passthrough? https://github.com/GiganticMinecraft/seichi_infra/blob/7a89624763eab97788abeff83dda2dc7a8a9b849/seichi-onp-k8s/manifests/seichi-kubernetes/apps/seichi-gateway/bungeecord/bungeecord-config-patch.yaml#L76

@weihao This is the default setting of BungeeCord. Anyways ping value (whether ping packet is passed through or not) difference is insignificant, RedisBungee is responsible for displaying the player count, and a MotD plugin is controlling motd so I didn't feel obliged to turn this on.

@weihao currently we still use on-prem vm redis and we are planning to migrate it to containers. Ref. GiganticMinecraft/seichi_infra#462

Regarding this setup, one mistake we have made is that we have unintentionally shared the single redis instance across two disjoint BungeeCord networks. However I don't see a reason why that could be causing the issue.

@weihao Thank you so much for the minimal self-contained setup! Once I have some more time this weekend I will setup a self-contained reproduction, and see what could be causing the issue.

[weihao]

Might be related to #88

The data source error should be resolved now