refactor: rename ScadaLink → ZB.MOM.WW.ScadaBridge (code + projects + namespaces)
Solution + 23 src projects + 26 test projects renamed; folders, csproj, namespaces, and ScadaLinkDbContext/ScadaBridgeDbContext class updated. ActorSystem "scadalink" → "scadabridge", Akka seed-node URLs migrated. SQL roles/logins, LDAP domains, CLI command name, and CLI config dir (~/.scadalink → ~/.scadabridge) also renamed. Build green; 5 Host.Tests fail awaiting SQL login rename in next commit. Pre-existing StaleTagMonitor timing flakes unchanged. Rename script committed at tools/rename-to-scadabridge.sh.
This commit is contained in:
Joseph Doherty
@@ -0,0 +1,211 @@
|
||||
using System.Collections.Concurrent;
|
||||
using ZB.MOM.WW.ScadaBridge.Commons.Messages.Health;
|
||||
using ZB.MOM.WW.ScadaBridge.Commons.Types;
|
||||
using ZB.MOM.WW.ScadaBridge.Commons.Types.Enums;
|
||||
|
||||
namespace ZB.MOM.WW.ScadaBridge.HealthMonitoring;
|
||||
|
||||
/// <summary>
|
||||
/// Collects health metrics from all site subsystems.
|
||||
/// Thread-safe: counters use Interlocked operations, connection/tag data uses ConcurrentDictionary.
|
||||
/// </summary>
|
||||
public class SiteHealthCollector : ISiteHealthCollector
|
||||
{
|
||||
private int _scriptErrorCount;
|
||||
private int _alarmErrorCount;
|
||||
private int _deadLetterCount;
|
||||
private int _siteAuditWriteFailures;
|
||||
private int _auditRedactionFailures;
|
||||
private volatile SiteAuditBacklogSnapshot? _siteAuditBacklog;
|
||||
private readonly ConcurrentDictionary<string, ConnectionHealth> _connectionStatuses = new();
|
||||
private readonly ConcurrentDictionary<string, TagResolutionStatus> _tagResolutionCounts = new();
|
||||
private readonly ConcurrentDictionary<string, string> _connectionEndpoints = new();
|
||||
private readonly ConcurrentDictionary<string, TagQualityCounts> _tagQualityCounts = new();
|
||||
private IReadOnlyDictionary<string, int> _sfBufferDepths = new Dictionary<string, int>();
|
||||
private int _deployedInstanceCount, _enabledInstanceCount, _disabledInstanceCount;
|
||||
private int _parkedMessageCount;
|
||||
private volatile string _nodeHostname = "";
|
||||
private volatile IReadOnlyList<Commons.Messages.Health.NodeStatus>? _clusterNodes;
|
||||
private volatile bool _isActiveNode;
|
||||
private readonly TimeProvider _timeProvider;
|
||||
|
||||
/// <summary>
|
||||
/// Creates a collector. The <paramref name="timeProvider"/> stamps each
|
||||
/// report's timestamp; it defaults to <see cref="TimeProvider.System"/> and
|
||||
/// is injectable so the report timestamp is deterministically testable —
|
||||
/// consistent with the rest of the module's time-dependent classes.
|
||||
/// </summary>
|
||||
/// <param name="timeProvider">Optional custom time provider; defaults to system time.</param>
|
||||
public SiteHealthCollector(TimeProvider? timeProvider = null)
|
||||
{
|
||||
_timeProvider = timeProvider ?? TimeProvider.System;
|
||||
}
|
||||
|
||||
/// <inheritdoc />
|
||||
public void IncrementScriptError()
|
||||
{
|
||||
Interlocked.Increment(ref _scriptErrorCount);
|
||||
}
|
||||
|
||||
/// <inheritdoc />
|
||||
public void IncrementAlarmError()
|
||||
{
|
||||
Interlocked.Increment(ref _alarmErrorCount);
|
||||
}
|
||||
|
||||
/// <inheritdoc />
|
||||
public void IncrementDeadLetter()
|
||||
{
|
||||
Interlocked.Increment(ref _deadLetterCount);
|
||||
}
|
||||
|
||||
/// <inheritdoc />
|
||||
public void IncrementSiteAuditWriteFailures()
|
||||
{
|
||||
Interlocked.Increment(ref _siteAuditWriteFailures);
|
||||
}
|
||||
|
||||
/// <inheritdoc />
|
||||
public void IncrementAuditRedactionFailure()
|
||||
{
|
||||
Interlocked.Increment(ref _auditRedactionFailures);
|
||||
}
|
||||
|
||||
/// <inheritdoc />
|
||||
public void UpdateSiteAuditBacklog(SiteAuditBacklogSnapshot snapshot)
|
||||
{
|
||||
_siteAuditBacklog = snapshot ?? throw new ArgumentNullException(nameof(snapshot));
|
||||
}
|
||||
|
||||
/// <inheritdoc />
|
||||
public void UpdateConnectionHealth(string connectionName, ConnectionHealth health)
|
||||
{
|
||||
_connectionStatuses[connectionName] = health;
|
||||
}
|
||||
|
||||
/// <inheritdoc />
|
||||
public void RemoveConnection(string connectionName)
|
||||
{
|
||||
_connectionStatuses.TryRemove(connectionName, out _);
|
||||
_tagResolutionCounts.TryRemove(connectionName, out _);
|
||||
_connectionEndpoints.TryRemove(connectionName, out _);
|
||||
_tagQualityCounts.TryRemove(connectionName, out _);
|
||||
}
|
||||
|
||||
/// <inheritdoc />
|
||||
public void UpdateTagResolution(string connectionName, int totalSubscribed, int successfullyResolved)
|
||||
{
|
||||
_tagResolutionCounts[connectionName] = new TagResolutionStatus(totalSubscribed, successfullyResolved);
|
||||
}
|
||||
|
||||
/// <inheritdoc />
|
||||
public void UpdateConnectionEndpoint(string connectionName, string endpoint)
|
||||
{
|
||||
_connectionEndpoints[connectionName] = endpoint;
|
||||
}
|
||||
|
||||
/// <inheritdoc />
|
||||
public void UpdateTagQuality(string connectionName, int good, int bad, int uncertain)
|
||||
{
|
||||
_tagQualityCounts[connectionName] = new TagQualityCounts(good, bad, uncertain);
|
||||
}
|
||||
|
||||
/// <inheritdoc />
|
||||
public void SetParkedMessageCount(int count)
|
||||
{
|
||||
Interlocked.Exchange(ref _parkedMessageCount, count);
|
||||
}
|
||||
|
||||
/// <inheritdoc />
|
||||
public void SetNodeHostname(string hostname) => _nodeHostname = hostname;
|
||||
|
||||
/// <inheritdoc />
|
||||
public void SetClusterNodes(IReadOnlyList<Commons.Messages.Health.NodeStatus> nodes) => _clusterNodes = nodes;
|
||||
|
||||
/// <inheritdoc />
|
||||
public void SetStoreAndForwardDepths(IReadOnlyDictionary<string, int> depths)
|
||||
{
|
||||
_sfBufferDepths = depths;
|
||||
}
|
||||
|
||||
/// <inheritdoc />
|
||||
public void SetInstanceCounts(int deployed, int enabled, int disabled)
|
||||
{
|
||||
Interlocked.Exchange(ref _deployedInstanceCount, deployed);
|
||||
Interlocked.Exchange(ref _enabledInstanceCount, enabled);
|
||||
Interlocked.Exchange(ref _disabledInstanceCount, disabled);
|
||||
}
|
||||
|
||||
/// <inheritdoc />
|
||||
public void SetActiveNode(bool isActive) => _isActiveNode = isActive;
|
||||
|
||||
/// <inheritdoc />
|
||||
public bool IsActiveNode => _isActiveNode;
|
||||
|
||||
/// <inheritdoc />
|
||||
public void AddIntervalCounters(
|
||||
int scriptErrors,
|
||||
int alarmErrors,
|
||||
int deadLetters,
|
||||
int siteAuditWriteFailures,
|
||||
int auditRedactionFailures)
|
||||
{
|
||||
// HealthMonitoring-017: each counter is restored atomically via
|
||||
// Interlocked.Add so an increment that arrived during the failed Send
|
||||
// (and therefore accumulated against the zero left by CollectReport's
|
||||
// Exchange) is correctly summed with the values being put back. No
|
||||
// ordering between the five Adds is required — they target independent
|
||||
// fields.
|
||||
if (scriptErrors != 0) Interlocked.Add(ref _scriptErrorCount, scriptErrors);
|
||||
if (alarmErrors != 0) Interlocked.Add(ref _alarmErrorCount, alarmErrors);
|
||||
if (deadLetters != 0) Interlocked.Add(ref _deadLetterCount, deadLetters);
|
||||
if (siteAuditWriteFailures != 0) Interlocked.Add(ref _siteAuditWriteFailures, siteAuditWriteFailures);
|
||||
if (auditRedactionFailures != 0) Interlocked.Add(ref _auditRedactionFailures, auditRedactionFailures);
|
||||
}
|
||||
|
||||
/// <inheritdoc />
|
||||
public SiteHealthReport CollectReport(string siteId)
|
||||
{
|
||||
// Atomically read and reset the counters
|
||||
var scriptErrors = Interlocked.Exchange(ref _scriptErrorCount, 0);
|
||||
var alarmErrors = Interlocked.Exchange(ref _alarmErrorCount, 0);
|
||||
var deadLetters = Interlocked.Exchange(ref _deadLetterCount, 0);
|
||||
var siteAuditWriteFailures = Interlocked.Exchange(ref _siteAuditWriteFailures, 0);
|
||||
var auditRedactionFailures = Interlocked.Exchange(ref _auditRedactionFailures, 0);
|
||||
|
||||
// Snapshot current connection and tag resolution state
|
||||
var connectionStatuses = new Dictionary<string, ConnectionHealth>(_connectionStatuses);
|
||||
var tagResolution = new Dictionary<string, TagResolutionStatus>(_tagResolutionCounts);
|
||||
var connectionEndpoints = new Dictionary<string, string>(_connectionEndpoints);
|
||||
var tagQuality = new Dictionary<string, TagQualityCounts>(_tagQualityCounts);
|
||||
|
||||
// Snapshot current S&F buffer depths
|
||||
var sfBufferDepths = new Dictionary<string, int>(_sfBufferDepths);
|
||||
|
||||
// Determine node role from active/standby state
|
||||
var nodeRole = _isActiveNode ? "Active" : "Standby";
|
||||
|
||||
return new SiteHealthReport(
|
||||
SiteId: siteId,
|
||||
SequenceNumber: 0, // Caller (HealthReportSender) assigns the sequence number
|
||||
ReportTimestamp: _timeProvider.GetUtcNow(),
|
||||
DataConnectionStatuses: connectionStatuses,
|
||||
TagResolutionCounts: tagResolution,
|
||||
ScriptErrorCount: scriptErrors,
|
||||
AlarmEvaluationErrorCount: alarmErrors,
|
||||
StoreAndForwardBufferDepths: sfBufferDepths,
|
||||
DeadLetterCount: deadLetters,
|
||||
DeployedInstanceCount: _deployedInstanceCount,
|
||||
EnabledInstanceCount: _enabledInstanceCount,
|
||||
DisabledInstanceCount: _disabledInstanceCount,
|
||||
NodeRole: nodeRole,
|
||||
NodeHostname: _nodeHostname,
|
||||
DataConnectionEndpoints: connectionEndpoints,
|
||||
DataConnectionTagQuality: tagQuality,
|
||||
ParkedMessageCount: Interlocked.CompareExchange(ref _parkedMessageCount, 0, 0),
|
||||
ClusterNodes: _clusterNodes?.ToList(),
|
||||
SiteAuditWriteFailures: siteAuditWriteFailures,
|
||||
AuditRedactionFailure: auditRedactionFailures,
|
||||
SiteAuditBacklog: _siteAuditBacklog);
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user