#Howdy. I'm managing a micro-mono Go
frigid notch
I'm trying to figure out if DaggerCloud caches the image layers for all of our build steps so that we can (a) trigger dagger from our CI in GitHub Actions and (b) reuse the cache for subsequent builds, say, in a given PR.
@quartz dust Would you please tell me about the cache offering from DaggerCloud (i.e. memory limit, how it's managed, anything else relevant)?
perhaps there's an external cache source, like GCP's Artifact Registry, that we can leverage. I don't know, so I'm just spit-balling here.
quartz dust
<@920499459484418068> Would you please tell me about the cache offering from Dag...
If you're planning to use ephemeral runners on GitHub actions, the best way to get a reusable cache is to pair Dagger with Depot. You can learn more about that here: https://dagger.io/blog/dagger-and-depot
Right now Dagger Cloud is an observability product that gives you deep tracing of your dagger pipeline regardless of where it runs.
frigid notch
Ah, so Dagger Cloud won't solve our caching problem, but it will show us what and how things run. Depot, on the other hand, will store the layers generated by Dagger and cache them. Did I get that right?
quartz dust
Ah, so Dagger Cloud won't solve our caching problem, but it will show us what an...
Yeah exactly!
It will show you what and how things run but also its the answer for this:
achieve build step parity in both dev workstations and CI workflows
frigid notch
Thanks Lev. I think it should be easy enough to create 1 or more reusable Dagger modules and then use them to build our applications. Finally, I can then plug things together in CI w/ Depot + Dagger Cloud