What it does
One panel. Read-only by default; every mutating action asks first, and the destructive ones make you type the name.
Cluster explorer
Jobs with real health — a running one with missing allocs is degraded — allocations, tasks, nodes, deployments. Multi-cluster, one switch.
Plan diff repo↔running
Open a .nomad/.hcl and see the diff against the running job before deploying. Server-side HCL parsing.
Streaming logs
stdout/stderr of any task in a dedicated Output channel — streaming, not polling. Several streams side by side.
Incident bundle
One click on a failed alloc: an incidents/ folder with a report (event timeline) and the tail of the logs attached.
Cluster snapshot
A markdown health report: problems on top, full table below. Perfect for the morning check.
Safe by default
ACL tokens from env vars only, never in settings/logs. A warning if a token would travel in cleartext over http.
Deep dive
The questions you actually ask during an incident — answered without leaving the editor.
Version history & revert
What did the last deploy change? Diff any two versions, then roll back with the plan preview in front of you. Read more →
Placement diagnostics
Why won't it schedule? The scheduler's counters as sentences: constraint, exhausted memory, empty datacenter, quota. Read more →
Usage vs requested
Live CPU and memory against the spec. Flags the task about to be OOM-killed and the reservation nobody uses. Read more →
Node drain & eligibility
Make a node ineligible before maintenance, drain it with a deadline, watch the remaining allocations. Read more →
Grep across allocations
Which instance logged that error? One search over stdout+stderr of every allocation, in parallel. Read more →
Drift between clusters
Same job on two clusters, field by field — plus a job × cluster matrix of docker images. Read more →
Documentation
Generated from docs/GUIDE.md in the repo, so it never drifts from the extension.
Setup
Clusters, namespaces and the ACL token rule: the token is never stored in your settings.
Recipes
Before a deploy · morning check · a failing allocation · node maintenance · is prod really like dev?
Command reference
All 24 commands, where each one lives and what it does.
Settings reference
Every nomadLens.* setting with type, default and meaning.
Troubleshooting
403s, timeouts, empty lists, logs that will not start — symptom to remedy.
Security
What is read-only, what asks for confirmation, and how tokens are handled.
Installation
From the Marketplace (publisher allannava95) or from a .vsix.
code --install-extension nomad-lens-<version>.vsix
Then configure your clusters in settings.json — the token is read from an env var, never stored:
"nomadLens.clusters": [
{ "name": "dev", "address": "http://nomad-dev.example:4646" },
{ "name": "prod", "address": "https://nomad.example:4646",
"namespace": "default", "tokenEnv": "NOMAD_TOKEN_PROD" }
]
The full documentation covers every command and setting, plus recipes for a deploy, the morning check, a failing allocation and node maintenance.