[20:12:52] Those redis errors seem to have fixed themselves. Unless someone looked into them. [20:13:49] that wasn't anything to do with redis, the cause was broken RCStream by WMF [20:14:22] And they fixed it I take it? [20:14:41] 'DEBUG XmlRcs FATAL: redis is empty for 10 seconds': perhaps this message should be changed. [20:14:47] I have no idea but it's probably more stable now [20:15:05] The bot needs to be reloaded anyway, to use those new translations. [20:15:24] https://wikitech.wikimedia.org/wiki/XmlRcs explains that a bit more in depth [20:15:55] * Matthew_ reads [20:15:58] the thing is there are 2 parts that XmlRcs consist from, the part behind redis is what reports the error and this part has no idea why redis is empty [20:16:20] it just knows it's empty for more than 10 seconds and that's clearly wrong [20:16:35] the other part uploads stuff to redis from RCStream [20:16:50] Hm. [20:16:53] RCfeed perhaps [20:17:48] [20:18:08] Daemon sends randomly these messages to verify if client is still online, if you fail to reply with pong you may get disconnected within a minute (note: the reply doesn't need to be "pong" the last_response time gets reset on any input) [20:18:25] Is there a possibility that it's not doing that?