Phase 3B: Site I/O & Observability — Communication, DCL, Script/Alarm actors, Health, Event Logging

Communication Layer (WP-1–5):
- 8 message patterns with correlation IDs, per-pattern timeouts
- Central/Site communication actors, transport heartbeat config
- Connection failure handling (no central buffering, debug streams killed)

Data Connection Layer (WP-6–14, WP-34):
- Connection actor with Become/Stash lifecycle (Connecting/Connected/Reconnecting)
- OPC UA + LmxProxy adapters behind IDataConnection
- Auto-reconnect, bad quality propagation, transparent re-subscribe
- Write-back, tag path resolution with retry, health reporting
- Protocol extensibility via DataConnectionFactory

Site Runtime (WP-15–25, WP-32–33):
- ScriptActor/ScriptExecutionActor (triggers, concurrent execution, blocking I/O dispatcher)
- AlarmActor/AlarmExecutionActor (ValueMatch/RangeViolation/RateOfChange, in-memory state)
- SharedScriptLibrary (inline execution), ScriptRuntimeContext (API)
- ScriptCompilationService (Roslyn, forbidden API enforcement, execution timeout)
- Recursion limit (default 10), call direction enforcement
- SiteStreamManager (per-subscriber bounded buffers, fire-and-forget)
- Debug view backend (snapshot + stream), concurrency serialization
- Local artifact storage (4 SQLite tables)

Health Monitoring (WP-26–28):
- SiteHealthCollector (thread-safe counters, connection state)
- HealthReportSender (30s interval, monotonic sequence numbers)
- CentralHealthAggregator (offline detection 60s, online recovery)

Site Event Logging (WP-29–31):
- SiteEventLogger (SQLite, 6 event categories, ISO 8601 UTC)
- EventLogPurgeService (30-day retention, 1GB cap)
- EventLogQueryService (filters, keyword search, keyset pagination)

541 tests pass, zero warnings.
This commit is contained in:
Joseph Doherty
2026-03-16 20:57:25 -04:00
parent a3bf0c43f3
commit 389f5a0378
97 changed files with 8308 additions and 127 deletions

View File

@@ -0,0 +1,126 @@
using Akka.Actor;
using Microsoft.CodeAnalysis.Scripting;
using Microsoft.Extensions.Logging;
using ScadaLink.Commons.Messages.ScriptExecution;
using ScadaLink.SiteRuntime.Scripts;
namespace ScadaLink.SiteRuntime.Actors;
/// <summary>
/// WP-15: Script Execution Actor -- short-lived child of Script Actor.
/// Receives compiled code, params, Instance Actor ref, and call depth.
/// Runs on a dedicated blocking I/O dispatcher.
/// Executes the script via Script Runtime API, returns result, then stops.
///
/// WP-32: Script failures are logged but do not disable the script.
/// Supervision: Stop on unhandled exception (parent ScriptActor decides).
/// </summary>
public class ScriptExecutionActor : ReceiveActor
{
public ScriptExecutionActor(
string scriptName,
string instanceName,
Script<object?> compiledScript,
IReadOnlyDictionary<string, object?>? parameters,
int callDepth,
IActorRef instanceActor,
SharedScriptLibrary sharedScriptLibrary,
SiteRuntimeOptions options,
IActorRef replyTo,
string correlationId,
ILogger logger)
{
// Immediately begin execution
var self = Self;
var parent = Context.Parent;
ExecuteScript(
scriptName, instanceName, compiledScript, parameters, callDepth,
instanceActor, sharedScriptLibrary, options, replyTo, correlationId,
self, parent, logger);
}
private static void ExecuteScript(
string scriptName,
string instanceName,
Script<object?> compiledScript,
IReadOnlyDictionary<string, object?>? parameters,
int callDepth,
IActorRef instanceActor,
SharedScriptLibrary sharedScriptLibrary,
SiteRuntimeOptions options,
IActorRef replyTo,
string correlationId,
IActorRef self,
IActorRef parent,
ILogger logger)
{
var timeout = TimeSpan.FromSeconds(options.ScriptExecutionTimeoutSeconds);
// CTS must be created inside the async lambda so it outlives this method
_ = Task.Run(async () =>
{
using var cts = new CancellationTokenSource(timeout);
try
{
var context = new ScriptRuntimeContext(
instanceActor,
self,
sharedScriptLibrary,
callDepth,
options.MaxScriptCallDepth,
timeout,
instanceName,
logger);
var globals = new ScriptGlobals
{
Instance = context,
Parameters = parameters ?? new Dictionary<string, object?>(),
CancellationToken = cts.Token
};
var state = await compiledScript.RunAsync(globals, cts.Token);
// Send result to requester if this was an Ask-based call
if (!replyTo.IsNobody())
{
replyTo.Tell(new ScriptCallResult(correlationId, true, state.ReturnValue, null));
}
// Notify parent of completion
parent.Tell(new ScriptActor.ScriptExecutionCompleted(scriptName, true, null));
}
catch (OperationCanceledException)
{
var errorMsg = $"Script '{scriptName}' on instance '{instanceName}' timed out after {timeout.TotalSeconds}s";
logger.LogWarning(errorMsg);
if (!replyTo.IsNobody())
{
replyTo.Tell(new ScriptCallResult(correlationId, false, null, errorMsg));
}
parent.Tell(new ScriptActor.ScriptExecutionCompleted(scriptName, false, errorMsg));
}
catch (Exception ex)
{
// WP-32: Failures logged to site event log; script NOT disabled after failure
var errorMsg = $"Script '{scriptName}' on instance '{instanceName}' failed: {ex.Message}";
logger.LogError(ex, "Script execution failed: {Script} on {Instance}", scriptName, instanceName);
if (!replyTo.IsNobody())
{
replyTo.Tell(new ScriptCallResult(correlationId, false, null, errorMsg));
}
parent.Tell(new ScriptActor.ScriptExecutionCompleted(scriptName, false, errorMsg));
}
finally
{
// Stop self after execution completes
self.Tell(PoisonPill.Instance);
}
});
}
}