Digital Ocean deployment fails when using current required resources and resources aren't enforced about charts-clickhouse HOT 6 CLOSED

Do we need 1000Mi for migrations really (looks like it came from the initial setup https://github.com/PostHog/vpc/commit/e26694eb9f4ec4821e8d89d27feca9c7b3ae7b7e)? Would 500Mi or 250Mi be good enough? That would longer term let us allow smaller deployments.

from charts-clickhouse.

We should be able to get by with 250 Mi no problem for migrations. It's just python calling out to clickhouse to run sql. Nothing actually is processed on migration side

from charts-clickhouse.

Give 500Mi or 250Mi a try and see if it works! 🤞

from charts-clickhouse.

Can totally try for some random recent migrations, but I was concerned that there might be some migrations that require more resources than others & I haven't had enough exposure to even know which kind could require more memory.

from charts-clickhouse.

Created an issue for resource requirements enforcement digitalocean/marketplace-kubernetes#241

from charts-clickhouse.

Closing this as the migrations part was addressed in #60 & 1,2 are really tied together with the same issue that I linked already (& we can't do anything here, but it's also less important now as the default cluster pre-filled values work now & we've clarified sizing in our docs).

from charts-clickhouse.