What are the OSX designers smoking? The roundness in Tahoe is so bad. Was this is even QA’ed?

Is there really no way of changing the max body size on a GKE Ingress? #gcp #gke #413 #entitytoolarge

Xandr down?

Back to using @ScyllaDB :) For the occasion, I gave a bit of love to marina my CQL Erlang client! github.com/lpgauth/marina/re…

Where do people buy their SSL certificates from? Ideally wildcard cert good for one year.

Big fan of all the managed DBs that offer 300$ credit to test out their offering. Much easier than setting up all the DBs!

I love HTTP 1.1 but we’re spending stupid amount on egress with just “204 No Content”. Wish it was a bit more bandwidth friendly.

Preference for handling money in the DB: 1) use decimal so there’s no loss of precision 2) multiply everything by 100 and use integers everywhere 1 or 2?

Still fighting with GKE Ingress / routing. For some reason I end up with 200K TCP SYN_SENT and it fills up my conntrack. I’ve already lowered the nf_conntrack_tcp_timeout_syn_sent to 60… I low can I go?

There goes 6 hours of debugging: “nf_conntrack: table full, dropping packet”

Is there an easy way to get actual backend latency in GKE hone using a global load balancer? I don’t care much about the routing in the internal Google network, only the last hop to my backend…

Let’s see if switching to a gke-l7-regional-external-managed is better. Otherwise back to setting up my own proxying.

Coming from a metal environment, I’m not a big fan of black box load balancers.

Is @github really charging Github Actions by the minute instead of seconds?What a scam.

Wow, OrbStack is so much better than Docker for Mac it’s not even funny. I might just stop using a remote VPS for my docker needs.

Never really got the chance to play with Terraform until now. It hasn’t been painless, but somewhat enjoyable so far.

IaC in the application repo or in a central infra repo?