chore: persist template import terraform plan in postgres #17012
Conversation
|
The purpose of this is not widely known, so just dropping a note. The output of The phasing in of this new feature is that it will live alongside the existing flow for some time. So this new table will be |
There was a problem hiding this comment.
This is all looking good. I am wondering if in the future we will need to implement a way to garbage collect up this data.
We might be able to do some trimming in the future, but keeping the raw plans gives us the most flexibility in the future.
EDIT: We store the original TF too, so versions are already bloated
Adds plumbing to persist terraform plan data from template import provisioner builds.
After some deliberation, I went with storing what I have called the
cached_planin a separate table because these values are incredibly large, thetemplate_versionstable already has a lot of columns, and a lot of queries usetemplate_version.*to get those many columns. Accidentally including thecached_plancolumn in those queries would be very undesirable, and this value mostly exists for future use on the backend. I don't think this is the sort of thing we would ever actually want to send to a client, except maybe a page for debugging the current state.This also means that we can track it's
updated_attime separately quite easily, since we're planning on letting it be "refreshed".Closes #17050