Quickstart
Run Lembos on your laptop with one command — a local k3s cluster, the chart, and Grafana over it.
One command puts a complete Lembos on your machine: a k3s cluster in Docker with the Helm chart installed
into it, one pod per service and Temporal in the release, PostgreSQL, Redis and a mail catcher beside it, and
Grafana over its traces, logs and metrics. It is served under *.dev.localhost with the developer
certificate your machine already trusts, so there is nothing to import.
It is the chart a production install uses, observed the way a production install would be — the fastest way to see the Portal with your own eyes, and to see what running it looks like.
What you need
-
Docker with about 4 CPUs and 8 GiB of memory to spare
-
The .NET 10 SDK
-
The Aspire CLI, with its developer certificate trusted:
aspire certs trust -
Ports 80, 443 and 3000 free — the ingress publishes on 80 and 443, Grafana on 3000
-
kubectl, optionally, for looking inside the cluster
1. Start
git clone https://github.com/lembosproj/deploy && cd deploy
aspire startThe Aspire dashboard's link is printed. The first start takes several minutes, most of it pulling images;
the lembos resource finishes once every pod is ready. Starting again later is quicker: the cluster keeps
its data, and each start upgrades the release to the chart and values in the checkout.
2. Sign in
Open https://portal.dev.localhost and sign in as [email protected]. The
password was generated for you:
source scripts/k3s-env.sh # points kubectl at the cluster, and adds `k` for the lembos namespace
lembos-admin-password| What | Where |
|---|---|
| Portal | https://portal.dev.localhost |
| Sign-in | https://identity.dev.localhost |
| GraphQL | https://orchestrator.dev.localhost/graphql |
| Workflows | https://temporal.dev.localhost |
| Grafana | http://localhost:3000 |
the mailpit resource in the Aspire dashboard |
Invitations, confirmations and password resets land in the mail catcher; nothing leaves your machine.
Want something to look at?
The catalog starts empty. To browse a sample organisation — an invented retailer with teams, systems,
components, APIs, blueprints and environments — set orchestrator.sampleData: true in
aspire/Lembos.Deploy.AppHost/values.aspire.yaml, then aspire stop and aspire start.
The sample is for reading. Its cloud accounts point at a local AWS emulator the quickstart does not run, so checking them or provisioning its environments fails here.
3. Watch it run
The services send their telemetry to Grafana, and three dashboards open on it:
| Dashboard | What it answers |
|---|---|
| Lembos — Platform | Is the platform healthy? Requests, errors and latency per service, pods, logs, traces |
| Lembos — Service | Where is one service spending its time? |
| Lembos — Temporal | Is Temporal keeping up with the workflows Lembos gives it? |
Metrics arrive once a minute, so a fresh start takes a couple of minutes to fill in. The dashboards ship with the chart, so the same ones work against the Grafana in your own cluster.
Clean up
aspire stopThe cluster and its dependencies are persistent, so they keep running and keep their data. To remove them for a fresh start, follow Tear down in the chart's README.
If something does not come up, the chart's README has what to look at.
Next
- Walkthrough — from this empty quickstart to a service running with its database and cache.
- Concepts — what you are looking at.
- Install — the same chart in a cluster other people use.