RoyakRoyak connecting…

This page runs on Royak.

Deployed by a git push, built and served by Royak's own ingress. Every number below is this cluster's real, live state — polled from Royak's API every 2 seconds.

This response served by pod 1d59e6e0aa1e · now — refresh and Royak's ingress round-robins you to a different replica. Not a static page.

A chaos monkey is loose on purpose. Royak kills a random pod every ~20 seconds so you can watch it notice and self-heal — every time, automatically. Nothing here is broken; the pods dying and coming back is the demo. Every number below is this live 2-node cluster.
In front of Royak sits NeuroCaddy — our own Rust web server + TLS edge. It terminates HTTPS (Let's Encrypt), load-balances, and runs a small neural network that learns this site's traffic patterns to flag anomalies and shape routing decisions. Edge, orchestrator and cross-node mesh — the whole stack is homegrown.
⚡ Benchmarked live — the whole path
622
req / sec
98.3%
success @ 10k req
143ms
p50 latency
200
concurrent
NeuroCaddy → Royak ingress → round-robin → 8 pods across 2 nodes · cluster: 2× Oracle Cloud Ampere, 1 vCPU / 1 GB each, cross-provider private mesh (AES-256-GCM)
Pods running
Deployments
Nodes
Services
Namespaces
mTLS certs
Reconcile ticks
Cluster uptime

Resilience — self-healing across restarts

0
Disruptions
0
Recovered
Avg recovery
Last recovery
Stable for

Pods over time

mTLS certs issued

Per-pod CPU

Per-pod memory

Live orchestrator activity watching…

waiting for Royak to reconcile…
Powered by Royak — a swarm of containers in one ~7.3 MB binary · github.com/evotech-bg/royak