daniellavoie opened a new issue #6971: URL: https://github.com/apache/incubator-pinot/issues/6971
# Context Been brainstorming with @mayankshriv about my observations regarding Pinot stability. From running clusters in production and troubleshooting people in our community Slack, my takeaway is that it’s very easy to crash a Pinot Server if you a create tables for which your system is not resourced properly. OOM, segments beings too big or too numerous are all correlated to your input dataset and available compute resources. It is very easy to shoot yourself in the foot and blow up your own cluster. I'm opening this issue to brainstorm with the rest of the community and see how people think we could make Pinot more reliable from user provided settings. # Proposal Make Pinot self aware of its resource availability, introduce circuit-breakers and fail-fast mechanism to prevent such things and have Pinot operate within available resources. This is certainly easier said than done. Making Pinot more reliable to user provided settings could be acheive with 2 combinable strategies: 1. Introduce an auto-tuned ingestion mode that can ingestion settings (batch size, segment size, etc) based on available resources. 2. Introduce circuit breakers and fail-fast mechanism to prevent Pinot server to crash from OOM or FS overflows but still serve queries. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org For additional commands, e-mail: commits-h...@pinot.apache.org
