as a bigCorp engineer I love k8s, but we have a dedicated team running it

for my solo efforts I prefer giving a dockerfile to a platform like render or heroku and letting them deal with config

3. your career is a pie-eating contest where the prize for winning is more pie

the things people see you do are the things they'll ask you to do. if you want to get asked to do certain types of work, make sure people see you doing that type of work

that's insane

that's the same setup I ended up with hah

oh it's client side rendered, the timings above are just for loading graphql data

they're both in us-east-1, should be 20ish ms at most

Next step: move the dashboard over to the existing public REST API.

Historically it's been hard to dogfood the public REST API since the whole dash uses GraphQL, so looking forward to that changing.

I've managed to speed it up by ~80% with one simple trick: stop relying on GraphQL.

Instead of fetching data resolver-by-resolver, now I fetch all of the data in one chonker of an SQL query.

I just use Trello and the check list feature

yep - own hardware, own time, own IP

#1*

this happens literally every time for me, the worst

when i can't breathe while sitting en terrasse with my pinte de bière