devinay commented on pull request #5961: URL: https://github.com/apache/incubator-pinot/pull/5961#issuecomment-686246344
This is good. I had a few additional comments: It would probably make sense to have 3 types of installation : frontend, worker and scheduler : So users should be able to install these 3 kinds of nodes either by changing something in values.yaml or by passing params to helm. This is convenient as each of these 3 node types could live in different VPC's or be provisioned in different env's. Could the mysql initialization be treated as an independent concern -- because its a one-time thing. It may not be a bad thing to assume that the db is already available and initialized and not treat DB upgrades via helm at all. DB install / backup / schema migration is an independent and mature concern with its own tools like liquibase. In theory, thirdeye should be able to work with any database, maybe even NoSQL ones. The bulk of operation load (IMHO) comes because we want to scale or change the config of some existing cluster, we want to increase the size of the back end, add new config variables, override old ones, support new datasources or detection algos. This would help upgrade / scale up a cluster to keep pace with code changes and load requirements. ---------------------------------------------------------------- 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
