Hello team! I am looking into Pinot for a realtime gaming platform, basically we have a stream of events we want to send to kafka from the frontend, and pinot ideally should power the stats the players see like leaderboards and profiles. Also even realtime stats while you are playing, like for example player killed, headshot, bodyshot, movements even, so all of this data will get ingested into pinot, either direct kafka, or socket with our backend that handles publishing to kafka topics. Additionally, at the same time players are constantly querying the leaderboard, tournament stats, from pinot, so its a 2 way street kind of. Of course I will use redis to cache this data when possible, however in game it will be difficult because the state will constantly change. My concern is if pinot is well suited for this, from what ive read its really amazing and looks quite powerful for this use case. Seems you guys did an amazing job! To summarize I am exploring pinot for this realtime game tournament concept. Am i going in the right direction? Can Pinot handle such a heavy load?

There’s some missing info, e.g. number of events per second, player queries per second, and some general description of the types of queries. You also don’t specify max latency between an event being sent to Kafka, and the results being available via Pinot. But in general, yes it should work for your use case.

alright thanks @Ken Krugler ! will come back with more info, but you can expect a million events per second at scale 😅, I believe pinot clusters today on production are already capable of ingesting this many records per second through kafka. However i think this warrants a more detailed breakdown of the queries, events, and how they might grow at scale. Will definitely share this breakdown here on #CDRCA57FC, hope this discussion invites more people to use pinot in the gaming world 🍻