Magellan - world packet navigation
http://localhost:8080/nbi/swagger-ui.html
source
docker run -d -p 8888:8080 --name magellan obrienlabs/magellan-nbi:0.0.3-arm
start forwarding thread http://127.0.0.1:8888/nbi/forward/packet?dnsFrom=host.docker.internal&dnsTo=host.docker.internal&from=8889&to=8888&delay=1000
start/stop thread http://127.0.0.1:8888/nbi/forward/reset
- source URL
- target URL
- message
- timestamps
- node
- timestamp
HTTPS support is added via Google Cloud Run
source swagger: https://traffic-generation-source-vyua7q27tq-nn.a.run.app/nbi/swagger-ui.html#/application-service-controller/getHealthUsingGET_1 source: https://traffic-generation-source-vyua7q27tq-nn.a.run.app/nbi/api
target: https://traffic-generation-target-vyua7q27tq-pd.a.run.app/nbi/api
Traffic gen API:
source
curl -X GET "https://traffic-generation-source-vyua7q27tq-nn.a.run.app/nbi/forward/packet?delay=1000&dnsFrom=traffic-generation-source-vyua7q27tq-nn.a.run.app&dnsTo=traffic-generation-target-vyua7q27tq-pd.a.run.app&from=80&to=80" -H "accept: /"
target
2 copies of the traffic gen containers are currently running on cloud run until I move them over to the GKE cluster. The health check doubles as the target REST endpoint for now. Note: HTTPS termination is on the cloud run gateway - if we don't terminate on the API Gateway and/or the LB fronting the GKE cluster then I'll add a self signed to the app - for now we use the Google cert for https.
CICD Currently there is a CSR trigger on cloud build to run the maven build and build/push the container into artifact registry. The cloud run builds are static but can be switched to trigger. I expect we can add a trigger on the cloud deploy job once I have moved the container to GKE.
To view - open swagger or use direct curls or the browser
enter for example 120k 1ms traffic https://traffic-generation-source-vyua7q27tq-nn.a.run.app/nbi/forward/traffic?delay=1&dns=traffic-generation-target-vyua7q27tq-pd.a.run.app&iterations=120000&to=80
view the metrics on both source and target.
These requests have no real load on the backend (db marshall/unmarshall as part of orm) therefore 1k requests/sec only scales a smaller than normal 1g/1vCore container up to 3 instances.
Once we get backend load scaling should occur with over 1k req/sec traffic
curl -X GET "http://127.0.0.1:8080/nbi/forward/traffic?dns=127.0.0.1&to=8080&delay=1000&iterations=20"
curl -X GET "http://127.0.0.1:8080/nbi/forward/reset"
gcloud compute instances create-with-container instance-20240227-002215 --project=cuda-old --zone=us-central1-a --machine-type=e2-medium --network-interface=address=34.69.213.211,network-tier=PREMIUM,stack-type=IPV4_ONLY,subnet=default --maintenance-policy=MIGRATE --provisioning-model=STANDARD --service-account=196717963363-compute@developer.gserviceaccount.com --scopes=https://www.googleapis.com/auth/cloud-platform --image=projects/cos-cloud/global/images/cos-stable-109-17800-147-15 --boot-disk-size=10GB --boot-disk-type=pd-balanced --boot-disk-device-name=instance-20240227-002215 --container-image=obrienlabs/magellan-nbi:0.0.3-ia64 --container-restart-policy=always --no-shielded-secure-boot --shielded-vtpm --shielded-integrity-monitoring --labels=goog-ec-src=vm_add-gcloud,container-vm=cos-stable-109-17800-147-15
Just need to expose the ports
michael@instance-20240227-002215 ~ $ docker ps -a
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
d56ed6dbcdde obrienlabs/magellan-nbi:0.0.3-ia64 "java -Djava.securit…" About an hour ago Up About an hour klt-instance-20240227-002215-mdvq
michael@instance-20240227-002215 ~ $ docker exec -it d56ed6dbcdde bash
root@instance-20240227-002215:/# curl http://127.0.0.1:8080/nbi/api
{"id":1,"content":"PASS remoteAddr: 127.0.0.1 localAddr: 127.0.0.1 remoteHost: 127.0.0.1 serverName: 127.0.0.1"}root@instance-20240227-002215:/#
open up the firewall to instances not just tags
{
"allowed": [
{
"IPProtocol": "tcp",
"ports": [
"80"
]
}
],
"creationTimestamp": "2023-11-30T10:27:00.856-08:00",
"description": "",
"direction": "INGRESS",
"disabled": false,
"enableLogging": false,
"id": "1173403927618874651",
"kind": "compute#firewall",
"logConfig": {
"enable": false
},
"name": "default-allow-http",
"network": "projects/cuda-old/global/networks/default",
"priority": 1000,
"selfLink": "projects/cuda-old/global/firewalls/default-allow-http",
"sourceRanges": [
"0.0.0.0/0"
],
"targetTags": [
"http-server"
]
}
