Dry-Run Mode¶

Dry-run mode lets you evaluate the full rule pipeline for an action without actually executing it, recording state changes, or emitting audit records. This is essential for:

• Testing rule changes before deploying
• Debugging why an action was suppressed, rerouted, or throttled
• Building rule-authoring tools with instant feedback

How It Works¶

Pass ?dry_run=true as a query parameter on the dispatch endpoint:

POST /v1/dispatch?dry_run=true
POST /v1/dispatch/batch?dry_run=true

The gateway will:

1. Evaluate all rules (including CEL expressions) to produce a verdict
2. Run the LLM guardrail check (if configured)
3. Return a DryRun outcome describing what would happen

The gateway will skip:

• Distributed lock acquisition
• Provider execution
• State mutations (dedup keys, group state, chain state, etc.)
• Audit record emission
• Rate-limit counter updates

Response Format¶

The DryRun outcome contains:

Example: Action would be allowed¶

curl -X POST 'http://localhost:8080/v1/dispatch?dry_run=true' \
  -H 'Content-Type: application/json' \
  -d '{
    "namespace": "notifications",
    "tenant": "tenant-1",
    "provider": "email",
    "action_type": "send_email",
    "payload": {"to": "user@example.com"}
  }'

{
  "DryRun": {
    "verdict": "allow",
    "matched_rule": null,
    "would_be_provider": "email"
  }
}

Example: Action would be suppressed¶

{
  "DryRun": {
    "verdict": "suppress",
    "matched_rule": "block-spam",
    "would_be_provider": "email"
  }
}

Example: Action would be rerouted¶

{
  "DryRun": {
    "verdict": "reroute",
    "matched_rule": "reroute-urgent",
    "would_be_provider": "sms"
  }
}

Client SDK Usage¶

Rust¶

use acteon_client::ActeonClient;
use acteon_core::Action;

let client = ActeonClient::new("http://localhost:8080");
let action = Action::new("ns", "tenant", "email", "send", serde_json::json!({}));

let outcome = client.dispatch_dry_run(&action).await?;
println!("{:?}", outcome);

Python¶

from acteon_client import ActeonClient

client = ActeonClient("http://localhost:8080")
action = Action(namespace="ns", tenant="t1", provider="email",
                action_type="send", payload={})
outcome = client.dispatch(action, dry_run=True)
if outcome.is_dry_run():
    print(f"Verdict: {outcome.verdict_details['verdict']}")

Node.js / TypeScript¶

const outcome = await client.dispatch(action, { dryRun: true });
if (outcome.type === "dry_run") {
  console.log(`Verdict: ${outcome.verdict}`);
}

Go¶

outcome, err := client.DispatchDryRun(ctx, action)
if outcome.IsDryRun() {
    fmt.Println("Verdict:", outcome.Verdict)
}

Java¶

ActionOutcome outcome = client.dispatchDryRun(action);
if (outcome.isDryRun()) {
    System.out.println("Verdict: " + outcome.getVerdict());
}

Simulation Framework¶

The simulation harness also supports dry-run mode:

use acteon_simulation::prelude::*;
use acteon_core::Action;

let harness = SimulationHarness::start(
    SimulationConfig::builder()
        .nodes(1)
        .add_recording_provider("email")
        .add_rule_yaml(r#"
            rules:
              - name: block-spam
                priority: 1
                condition:
                  field: action.action_type
                  eq: "spam"
                action:
                  type: suppress
        "#)
        .build()
).await.unwrap();

let action = Action::new("ns", "t1", "email", "spam", serde_json::json!({}));
let outcome = harness.dispatch_dry_run(&action).await.unwrap();

outcome.assert_dry_run();
// Provider was NOT called
harness.provider("email").unwrap().assert_not_called();

Batch Dry-Run¶

Batch dispatch also supports dry-run. All actions in the batch are evaluated independently and none are executed:

curl -X POST 'http://localhost:8080/v1/dispatch/batch?dry_run=true' \
  -H 'Content-Type: application/json' \
  -d '[
    {"namespace": "ns", "tenant": "t1", "provider": "email", "action_type": "send", "payload": {}},
    {"namespace": "ns", "tenant": "t1", "provider": "email", "action_type": "spam", "payload": {}}
  ]'

[
  {"DryRun": {"verdict": "allow", "matched_rule": null, "would_be_provider": "email"}},
  {"DryRun": {"verdict": "suppress", "matched_rule": "block-spam", "would_be_provider": "email"}}
]