You can speed up image builds by making use "--cache-from". This is useful when 1) you already have a previous version of the image which you can pull 2) docker pull is faster than docker build 3) intermediate layers in the previous version can be reused. This is useful for speeding up builds in docker container builder

There is something called QoS for pods using which we can prevent unwanted rescheduling of pods

you can use GCP cloud shell and use it to run kubectl commands instead of having to create a k8s context on your local kubectl CLI tool. This is very useful for doing some adhoc debugging on a k8s cluster you have access to. Just login to google cloud using google auth, and then click on "Connect" next to the cluster you want to connect to, then use cloud shell.

We can leverage structs in ruby to encapsulate and hold internal class data into a struct object. Since struct saves us from writing attr_accessor methods or initializer method, we can easily leverage it when in need of a temporary data structure and need to have the data grouped.

If you are using materialize, you need to initialize the select element with material_select(). Otherwise the select box will not get rendered.

Elixir and Erlang has a limit on the number of atoms that can be created. We can view the atom limit using :erlang.system_info(:atom_limit). Atom's text value is stored in the atom table and this is not garbage collected. This explains why user inputs to phoenix routes are to be matched with strings rather than atoms.