Guides
This section collects end-to-end walkthroughs for common operator tasks. Each guide is self-contained: concrete YAML, commands, and what to verify. Conceptual detail and field reference are described in Control plane, Policy & routing, and Observability.
Prerequisites for most guides: Conduit installed (Install and run), a runnable baseline config (Minimal configuration), and — where noted — the control plane enabled with control.listen_address.
Primary lab configs for guides that ship a full runnable sample are also installed with the production package and tarball under /usr/share/doc/conduit/examples/ (or examples/ in the tarball). See that directory’s README.md for the map from path to guide.
After your first successful query (First query), work top-down for a typical path — reliability → policy → live change → observability → capacity — or jump to the task you need.
Routing and reliability
| Guide |
What you practice |
| Backend health |
Enable probes, watch a dead backend go down, practice conduitctl health drain/resume |
| DNS answer cache |
Enable memory or LMDB caching, hit/miss path, on_hit and metrics tradeoffs |
| Dual-stack forwarding |
Global and per-pool egress sources, rules, and Rhai overrides |
Policy
| Guide |
What you practice |
| Rule action order |
Soft vs hard drop, clear_drop, action-list order, request set_retry_pool stash |
| Declarative failover |
SERVFAIL / timeout retry to another pool or backend without Rhai |
| Rhai policy |
CSV blocklist drop with custom user metric; CSV-driven pool routing |
Control plane
| Guide |
What you practice |
| Control plane workflows |
Reload from disk, temporary conduitctl apply overlays, export, and when to restart |
Observability
| Guide |
What you practice |
| Dataplane runtime tuning |
Choosing sync vs split_io, sizing ingress/policy/I/O workers, the slot pool, and reuse_port |