cdxker cdxker

Run this, with a different tag name in hn_frontend/frontend

docker build -t trieve/hn-ui:21 . && docker push trieve/hn-ui:21
gcloud container clusters get-credentials --region us-west1-a hackernews
kubectl edit -n hackernews deployments/hackernews

You will be in a vim session and just edit the tag here

Using raw vms are great for quickly self-hosting many products. Self hosting sentry is 50 services managed in docker-compose. The docs's arent very clear, but there is a way to use s3 to host the filestorage. Updating the config was too hard for me to figure out and now its all being logged to storage on the vm itself. Our selfhosted Sentry has died many times because it ran out of storage. I didn't realize that before. When you run out of storage, you are no longer able to even ssh into the machine to diagnose the problem. Anyways... this is how we do it on gcp.

From the gcloud cli run.

most annoying thing about eks,,, you gotta build everything yourself

AWS_REGION=us-west-1
CLUSTER_NAME=my-cluster
AWS_ACCOUNT_ID="$(aws sts get-caller-identity --query "Account" --output text)"

## Creating IAM Roles
eksctl create iamserviceaccount \
  --region="${AWS_REGION}" \
  --name="ebs-csi-controller-sa" \

Tailscale is OIDC compliant now!!!

We self-hosted headscale, but wanted to move to their cloud product for higher reliablity.

The way that tailscale authenticates for an email like [email protected] is:

Makes a call to https://trieve.ai/.well-known/webfinger?resource=acct:[email protected]. The response should look like this

	# This file is automatically @generated by Cargo.
	# It is not intended for manual editing.
	version = 3

	[[package]]
	name = "actix-codec"
	version = "0.5.2"
	source = "registry+https://github.com/rust-lang/crates.io-index"
	checksum = "5f7b0a21988c1bf877cf4759ef5ddaac04c1c9fe808c9142ecb78ba97d97a28a"
	dependencies = [

	export DATASET_ID="<your-trieve-dataset>"
	export ORGANIZATION_ID="<your-trieve-organization>"
	export TRIEVE_API_KEY="<your-trieve-api-key>"
	export TRIEVE_URL="https://api.trieve.ai"

	PAYLOAD='{
	"crawl_options": {
	"allow_external_links" : false,
	"boost_titles": true,
	"exclude_paths": [],

	let dataset_id = $env.DATASET_ID
	let api_key = $env.API_KEY
	let clickhousePassword = $env.CLICKHOUSE_PASSWORD
	let headers = {
	"TR-Dataset": $dataset_id
	"Authorization": $api_key
	"X-API-Version": "V2"
	}

	def clickhouse-query [query: string] {

	BACKEND=$(kubectl get ing <ingress-name> -o json \| jq -j '.metadata.annotations."ingress.kubernetes.io/backends"' \| jq -j 'keys[0]')
	gcloud compute backend-services update $BACKEND --enable-cdn