Prometheus, in your pocket.
Your Kubernetes metrics, native on your phone. From a locked screen to a root cause in under a minute — no laptop, no browser tabs, no VPN dance at 2am.
From page to root cause in under a minute
No YAML on the phone. No browser tabs. No waiting for a laptop to wake up. Your dashboards are already authored in git — the phone just reads them.
The alert wakes you, not a laptop
It is 2:14am. The push carries the numbers already — you know it is real before you are upright.
Errors climbing, latency shifted
One tap in: the counters and the status breakdown show the shape of it, and roughly when it started.
The node that is actually on fire
CPU and memory against requests and limits narrow it to one host. Most people are still opening a browser.
Fewer features. Stronger opinions.
Made for the phone in your hand
Not a desktop dashboard shrunk down. Built for the two minutes after you get paged, wherever you are.
Dashboards live in your repo
You write them down and commit them instead of dragging panels in a browser. Reviews, history and rollback come for free.
Prometheus, and only Prometheus
No logs, no traces, no long list of half-finished integrations. Doing one thing means every screen can be good at it.
Your cluster decides who sees what
Permissions come from Kubernetes, not from us. There is no second list of users to keep in sync.
Your metrics never leave your side
Nothing ships to our cloud and your bill does not grow with your data. A small open-source connector dials out — nothing dials in. Switch it off and access ends that second.
Three steps, all outbound
Nothing we run ever opens a connection to you.
Sign up
Create your org at admin.vedavid.dev. Dashboards, teams and connectors are managed there — the phone is for reading.
admin.vedavid.dev
Install the connector
One open-source binary, written in Rust, deployed next to your Prometheus. Read the source before you run it.
helm install vedavid-connector
It registers itself
The connector registers with the relay over mTLS and holds the connection open. It always starts the connection — no incoming connections are accepted, ever.
connector → relay · 443 out
Copy a ready-made dashboard instead of writing one
Open-source dashboard files for common services. Drop one into your repo and it shows up on your phone.
Stuck? Curious? Just nosy? Come ask.
One Slack channel, open to everyone. Ask how the connector works, why a query looks odd, or whether we'd ever support your weird setup. The person who wrote the code usually answers — and dumb questions are our favourite kind.
does the connector need cluster-admin? my platform team just made a face
cluster-admin? it doesn't even get cluster-anything. zero RBAC. the kubelet hands it a ConfigMap and it phones home. that's the whole trick
it needs fewer permissions than your platform team's coffee machine. zero RBAC — dashboards arrive as a mounted file, and it only dials out
the RBAC file is 0 lines. i checked twice. happy to walk them through the other files
face has been un-made. thank you!