rabbitmqadmin-sdk

An async Rust SDK for the RabbitMQ Management HTTP API. It provides a strongly-typed client built on reqwest + serde covering monitoring, administration, and day-2 operations for RabbitMQ 3.12+ / 4.x brokers with the management plugin enabled (typically on port 15672).

Features

• Monitoring: cluster overview, nodes (memory/disk/fd metrics), queues, exchanges, connections, channels, consumers, message-rate stats.
• Administration: declare and delete queues, exchanges, bindings, vhosts, users, permissions, policies, operator policies, and runtime parameters.
• Operations: purge queues, publish and fetch messages over HTTP, close connections (with X-Reason), export/import definitions, six health checks, list alarms, and trigger queue-leadership rebalancing.
• Ergonomics: percent-encoding of vhost/queue/exchange segments handled for you, contextual NotFound errors, server-side pagination support, no_std-free async via tokio/reqwest.

Quickstart

Add the dependency and point a Client at your broker:

[dependencies]
rabbitmqadmin-sdk = "0.1"
tokio = { version = "1", features = ["rt-multi-thread", "macros"] }

use rabbitmqadmin_sdk::Client;

# async fn example() -> rabbitmqadmin_sdk::Result<()> {
let client = Client::new("http://localhost:15672", "guest", "guest")?;
let me = client.whoami().await?;
println!("logged in as {} (tags: {})", me.name, me.tags.join(","));

for q in client.list_queues().await? {
    println!("{}/{}: {} messages", q.vhost, q.name, q.messages.unwrap_or(0));
}
# Ok(())
# }

Use Client::builder(...) to supply a pre-configured reqwest::Client (custom TLS, timeouts, proxies) via .http_client(...).

Endpoint coverage

Area	Methods	Endpoints
Auth	whoami	GET /api/whoami
Overview	overview, cluster_name	GET /api/overview, `GET
Nodes	list_nodes, get_node	GET /api/nodes[/{name}]
Queues	list_queues (+_paged, _in_vhost), get_queue, declare_queue, delete_queue, purge_queue, queue_action, get_messages	`GET
Exchanges	list_exchanges (+_paged, _in_vhost), get_exchange, declare_exchange, delete_exchange, publish_to_exchange, list_exchange_bindings_source	`GET
Bindings	list_bindings (+_paged, _in_vhost), list_queue_bindings, list_bindings_between, bind, unbind	`GET
Vhosts	list_vhosts, get_vhost, create_vhost, delete_vhost, list_vhost_permissions	`GET
Users	list_users, get_user, create_user, delete_user, list_user_permissions	`GET
Permissions	set_permission, delete_permission	`PUT
Policies	list_policies (+_in_vhost), get_policy, set_policy, delete_policy; same set for operator policies	`GET
Parameters	list_parameters (+_in_vhost), get_parameter, set_parameter, delete_parameter	`GET
Connections	list_connections (+_paged), get_connection, close_connection, list_connection_channels	`GET
Channels	list_channels, get_channel	GET /api/channels[/{name}]
Consumers	list_consumers (+_in_vhost)	GET /api/consumers[/{vhost}]
Definitions	export_definitions (+_in_vhost), import_definitions	`GET
Health	health_check_alarms, health_check_local_alarms, health_check_port_listener, health_check_protocol_listener, health_check_node_is_quorum_critical, health_check_virtual_hosts	GET /api/health/checks/...
Ops	list_alarms, rebalance_queues	GET /api/alarms, POST /api/rebalance/queues

All path segments (vhosts, queue names, ...) are percent-encoded automatically — DEFAULT_VHOST (/) becomes %2F on the wire.

Examples

Two runnable examples live in examples/. Both default to http://localhost:15672 with guest/guest and honor the RABBITMQ_URL, RABBITMQ_USER, and RABBITMQ_PASS environment variables:

# Cluster snapshot: whoami, overview, per-node resources, top-10 queues by depth.
cargo run --example monitor_queues

# Admin workflow: declare exchange + queue, bind, publish, verify, purge, clean up.
cargo run --example declare_and_bind

Error handling

All fallible calls return rabbitmqadmin_sdk::Result<T> with a single Error enum:

• Error::NotFound(ctx) — the server returned 404; ctx identifies the missing resource (e.g. queue 'q1' in vhost '/').
• Error::Api { status, reason } — any other non-2xx response; reason carries the server's response body. Health checks report failure as 503, so a failed check surfaces as Error::Api { status: 503, .. } with the failure JSON in reason — match on it rather than treating it as an unexpected error.
• Error::Transport(..) — HTTP-level failures (DNS, TLS, timeouts, connection refused).
• Error::Deserialize { source, body } — the response didn't match the expected schema; the raw body is retained for debugging.
• Error::InvalidUrl(..) — malformed base URL (must be http:// or https://).

Pagination

Large list endpoints come in two flavors: list_x() returns the full Vec<T>, while list_x_paged(&PaginationQuery) returns a Paginated<T> envelope with items, page, page_size, total_count, and filtered_count. PaginationQuery supports page, page_size, name filtering, and use_regex. Note the server caps page_size at 500.

License

Apache-2.0