Dear @npfoss and Gather Dev. Team,
We have a concerning issue to bring to your attention regarding our bot’s frequent disconnection from the spaces of our clients whenever the space reaches its capacity. We believed that non-user bots would not have this issue, but upon investigation, we realized they are likely subject to the same limitation as regular users. We are unsure whether the bot got kicked out after the space reached the capacity limit or lost its connection for some other reason and couldn’t reconnect because of the limit.
We are reaching out to see if any of you have experienced similar issues with bots that are not logged in as players but connected to spaces getting kicked out whenever the space hits its user cap. This issue is impacting our ability to effectively utilize the Spaces of our Clients we are connected to with a bot.
We would greatly appreciate any help or insights you can provide us with, and we hope that revisions to the connection/reconnection logic can be made to resolve this issue.
Hey! Yeah sorry that was unclear – bots which have entered count like normal users towards the capacity limit. I think this mostly makes sense tbh since both the cost (to operate) and value people get out of the space are pretty correlated with the number of ‘players’ in the space, so if you want more than 10 total it’s not free. Sorry it wasn’t clear though!
Thanks.I’ll have to check the logs and get back to you.
We could not find any warnings for non bot users. At least not at the time when the 25th free seat, for example, is occupied by a user.The bot then lost the connection and could not reconnect.
We can say it that way because the bot in question tracks the number of users in the spaces live. Whenever the user limit is reached, the connection of our tracking bot gets terminated. so it seems like you can take the bot’s slot for not spawning. and when the limit is reached, no more connections are allowed.
hmm, that is strange… especially because we don’t enforce it base on connections, but entered players now (changed around the same time as the other capacity changes). so it’s odd that now you’re seeing what should have been the old behavior
once you have any logs about the closures it should be a lot easier to debug