galapas1/silver-enigma

ZTMesh Control Plane

ZTMesh Core Submodule

The "core" directory is a git submodule refering to ZTMesh Core. Use the following command to download the core submodule dependency.

git submodule update --init core

Web Interactions

http://127.0.0.1:9090/metrics http://0.0.0.0:8080/swagger

Deps

Install the following dependencies using Brew and Go install.

brew install prettier
brew install clang-format
brew install grc
go install github.com/segmentio/golines@latest
go install mvdan.cc/gofumpt@latest

Headscale Core

headscale last sync: bafb6791d3e61cc2fd9f283081885f29497150d2

Kafka machine updates

Adding the following lines to the config.yaml

#Change as appropriate
kafka:
  broker: localhost:9092

Or set the environment variable NINJA_KAFKA_BOOTSTRAP_SERVER; the later takes precidence.

For local debugging add the folllowing to a kafka.env file:

KAFKA_ENABLE_KRAFT=yes
KAFKA_CFG_PROCESS_ROLES=broker,controller
KAFKA_CFG_CONTROLLER_LISTENER_NAMES=CONTROLLER
KAFKA_CFG_LISTENERS=PLAINTEXT://:9092,CONTROLLER://:9093
KAFKA_CFG_LISTENER_SECURITY_PROTOCOL_MAP=CONTROLLER:PLAINTEXT,PLAINTEXT:PLAINTEXT
KAFKA_CFG_ADVERTISED_LISTENERS=PLAINTEXT://127.0.0.1:9092
KAFKA_BROKER_ID=1
[email protected]:9093
ALLOW_PLAINTEXT_LISTENER=yes

And run:

docker run \
          -d \
          --name optmate-kafka \
          --env-file kafka.env \
          -p 9092:9092 \
          bitnami/kafka:latest
#Note: this takes a few minutes to start up on my arm64 macbook.
# it might be bacause the bitnami kafka image is only for x86 and
# docker desktop has to run it via QEMU.

Then create the topics (you will need to do brew install kafka to get these tools):

for t in register update delete
do
   kafka-topics --bootstrap-server 127.0.0.1:9092 \
      --create --topic machine.$t
done

Then you can view the topics like so:

kafka-console-consumer --bootstrap-server 127.0.0.1:9092 --topic machine.register
kafka-console-consumer --bootstrap-server 127.0.0.1:9092 --topic machine.update
kafka-console-consumer --bootstrap-server 127.0.0.1:9092 --topic machine.delete

Authorization Callback

When a client registers it's machine with ninja panda a machine.register Kafka message is send out with the expectation that something will eventually POST back to /api/v1/machine/register/callback/{corrId} with the authentication results. Note that, this authentication entity and it interaction with the client is outside the purview of ninjapanda. The only thing ninjapanda has to do with it is to notify the client where the authentication URL is. This is the role of this in the config:

machine_auth_url: <URL OF OAUTH IDENTIFIER SERVICE>

This is overriden by the NINJA_MACHINE_AUTH_URL environment variable.

In-Memory or Redis Cache

In the authorization flow above, ninjapanda either uses an in-memory cache (default) or Redis. To configure the Redis support use the following environment variables. For more details on the semantics of the Redis configurations see this page.

CACHE_TYPE=redis
CACHE_ADDRESS=your_redis_server.example.com:6379
CACHE_PASSWORD=top_secret
CACHE_MAX_RETRIES=10
CACHE_CONN_MAX_IDLE_TIME=1h
CACHE_POOL_SIZE=40
CACHE_MIN_IDLE_CONNS=3

Alternate RELAY Servers

By default the RELAY servers used are the ones defined in the config file, in the relay.urls property. This will not be viable for every environment. The list of URLs can be overriden with an environment variable. NINJA_RELAY_FILE_URLS which should be a comma-separated list of URLs that contain a list of RELAY regions and servers.

NINJA_RELAY_FILE_URLS=https://resources.optm.com/relay-prod01.json,https://resources.optm.com/relay-prod02.json

Clean cache

go clean -modcache

Logging How-to

https://betterstack.com/community/guides/logging/zerolog/

Testing

grpcurl -plaintext 0.0.0.0:50443 ninjapanda.v1.NinjapandaService/CheckHealth