Hi, our pipeline with garden sometimes take over an hour, is there a way to speed things up for the production pipeline?

Hi @colossal-salesclerk-87522! My first suggestion would be to use the verbose log level and to turn on timestamps (by passing the

--log-level=verbose

and

--show-timestamps

CLI options). This should show you the full logs for the Docker builds, along with a timestamp for each log line. Hopefully, this will reveal where the builds are spending most of their time (e.g. fetching layers from a registry, or at certain steps in the Dockerfiles). If you're using in-cluster building via the

cluster-buildkit

or

kaniko

build modes (see https://docs.garden.io/reference/providers/kubernetes#providers-.buildmode), you should also be able to trigger your build pipeline from your dev machine. That way, you might be able to quickly iterate on identifying the performance bottlenecks in your builds without having to repeatedly re-trigger your CI pipeline. Those would be my initial suggestions. We'd be happy to look at this more closely with you if anything unexpected shows up!