Merge branch 'worktree-agent-ac58bf5116b74ff84' into arch-review-remediation
This commit is contained in:
@@ -30,6 +30,9 @@ public class EventLogPurgeService : BackgroundService
|
||||
/// <summary>Number of events deleted per cap-purge batch.</summary>
|
||||
private const int CapPurgeBatchSize = 1000;
|
||||
|
||||
/// <summary>Number of events deleted per retention-purge batch — same size as <see cref="CapPurgeBatchSize"/>.</summary>
|
||||
private const int RetentionPurgeBatchSize = 1000;
|
||||
|
||||
private readonly SiteEventLogger _eventLogger;
|
||||
private readonly SiteEventLogOptions _options;
|
||||
private readonly ILogger<EventLogPurgeService> _logger;
|
||||
@@ -118,21 +121,46 @@ public class EventLogPurgeService : BackgroundService
|
||||
}
|
||||
}
|
||||
|
||||
/// <summary>
|
||||
/// Deletes events older than the retention window, in bounded
|
||||
/// <see cref="RetentionPurgeBatchSize"/>-row slices per DELETE statement rather than
|
||||
/// one unbounded DELETE for the whole expired set — mirroring the batching already
|
||||
/// used a few lines below in <see cref="PurgeByStorageCap"/> (<see cref="CapPurgeBatchSize"/>).
|
||||
/// An expired backlog in the tens/hundreds of thousands of rows would otherwise hold a
|
||||
/// single SQLite write transaction (and the shared <c>_writeLock</c>) open for the
|
||||
/// whole delete, blocking every concurrent <see cref="SiteEventLogger.LogEventAsync"/>
|
||||
/// writer-loop flush for that entire duration.
|
||||
/// </summary>
|
||||
private void PurgeByRetention()
|
||||
{
|
||||
var cutoff = DateTimeOffset.UtcNow.AddDays(-_options.RetentionDays).ToString("o");
|
||||
var totalDeleted = 0;
|
||||
|
||||
var deleted = _eventLogger.WithConnection(connection =>
|
||||
while (true)
|
||||
{
|
||||
using var cmd = connection.CreateCommand();
|
||||
cmd.CommandText = "DELETE FROM site_events WHERE timestamp < $cutoff";
|
||||
cmd.Parameters.AddWithValue("$cutoff", cutoff);
|
||||
return cmd.ExecuteNonQuery();
|
||||
});
|
||||
var deleted = _eventLogger.WithConnection(connection =>
|
||||
{
|
||||
using var cmd = connection.CreateCommand();
|
||||
cmd.CommandText = $"""
|
||||
DELETE FROM site_events WHERE id IN (
|
||||
SELECT id FROM site_events
|
||||
WHERE timestamp < $cutoff
|
||||
LIMIT {RetentionPurgeBatchSize}
|
||||
)
|
||||
""";
|
||||
cmd.Parameters.AddWithValue("$cutoff", cutoff);
|
||||
return cmd.ExecuteNonQuery();
|
||||
});
|
||||
|
||||
if (deleted > 0)
|
||||
totalDeleted += deleted;
|
||||
|
||||
if (deleted < RetentionPurgeBatchSize)
|
||||
break;
|
||||
}
|
||||
|
||||
if (totalDeleted > 0)
|
||||
{
|
||||
_logger.LogInformation("Purged {Count} events older than {Days} days", deleted, _options.RetentionDays);
|
||||
_logger.LogInformation("Purged {Count} events older than {Days} days", totalDeleted, _options.RetentionDays);
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
@@ -34,12 +34,23 @@ namespace ZB.MOM.WW.ScadaBridge.SiteEventLogging;
|
||||
/// </remarks>
|
||||
public class SiteEventLogger : ISiteEventLogger, IDisposable
|
||||
{
|
||||
/// <summary>
|
||||
/// Maximum number of queued events committed together in a single transaction
|
||||
/// per writer-loop drain. Mirrors <c>SqliteAuditWriter</c>'s batch size
|
||||
/// (<c>ZB.MOM.WW.ScadaBridge.AuditLog/Site/SqliteAuditWriter.cs</c>,
|
||||
/// <c>ProcessWriteQueueAsync</c>/<c>FlushBatch</c>) — see the comment on
|
||||
/// <see cref="ProcessWriteQueueAsync"/> for why one transaction per drain
|
||||
/// replaces the previous one-transaction-per-event shape.
|
||||
/// </summary>
|
||||
internal const int WriteBatchSize = 256;
|
||||
|
||||
private readonly SqliteConnection _connection;
|
||||
private readonly ILogger<SiteEventLogger> _logger;
|
||||
private readonly object _writeLock = new();
|
||||
private readonly Channel<PendingEvent> _writeQueue;
|
||||
private readonly Task _writerLoop;
|
||||
private long _failedWriteCount;
|
||||
private long _committedBatchCount;
|
||||
private bool _disposed;
|
||||
|
||||
/// <summary>
|
||||
@@ -102,6 +113,14 @@ public class SiteEventLogger : ISiteEventLogger, IDisposable
|
||||
/// <inheritdoc />
|
||||
public long FailedWriteCount => Interlocked.Read(ref _failedWriteCount);
|
||||
|
||||
/// <summary>
|
||||
/// Number of write-batch transactions successfully committed since startup. Test-only
|
||||
/// observability hook (not part of <see cref="ISiteEventLogger"/>) proving the
|
||||
/// "drain up to <see cref="WriteBatchSize"/> queued events into one transaction" shape
|
||||
/// described on <see cref="ProcessWriteQueueAsync"/> — production code does not read it.
|
||||
/// </summary>
|
||||
internal long CommittedBatchCount => Interlocked.Read(ref _committedBatchCount);
|
||||
|
||||
/// <summary>
|
||||
/// Runs <paramref name="action"/> against the shared connection while holding the
|
||||
/// write lock, so purge / query / record callers on different threads never use
|
||||
@@ -212,54 +231,132 @@ public class SiteEventLogger : ISiteEventLogger, IDisposable
|
||||
return pending.Completion.Task;
|
||||
}
|
||||
|
||||
/// <summary>
|
||||
/// Drains up to <see cref="WriteBatchSize"/> queued events per iteration and commits
|
||||
/// them in a single transaction, instead of one transaction per event. Mirrors the
|
||||
/// reference shape in <c>SqliteAuditWriter.ProcessWriteQueueAsync</c>/<c>FlushBatch</c>
|
||||
/// (<c>ZB.MOM.WW.ScadaBridge.AuditLog/Site/SqliteAuditWriter.cs</c>): pull the first
|
||||
/// item with the blocking <c>ReadAllAsync</c> enumerator, then greedily <c>TryRead</c>
|
||||
/// whatever else is already queued (non-blocking) up to the batch cap, so a burst of
|
||||
/// concurrent <see cref="LogEventAsync"/> callers amortises SQLite's per-transaction
|
||||
/// fsync/journal overhead across the whole burst instead of paying it per row.
|
||||
/// </summary>
|
||||
private async Task ProcessWriteQueueAsync()
|
||||
{
|
||||
await foreach (var pending in _writeQueue.Reader.ReadAllAsync().ConfigureAwait(false))
|
||||
var batch = new List<PendingEvent>(WriteBatchSize);
|
||||
|
||||
await foreach (var first in _writeQueue.Reader.ReadAllAsync().ConfigureAwait(false))
|
||||
{
|
||||
try
|
||||
batch.Clear();
|
||||
batch.Add(first);
|
||||
|
||||
while (batch.Count < WriteBatchSize && _writeQueue.Reader.TryRead(out var next))
|
||||
{
|
||||
var written = WithConnection(connection =>
|
||||
batch.Add(next);
|
||||
}
|
||||
|
||||
FlushBatch(batch);
|
||||
}
|
||||
}
|
||||
|
||||
/// <summary>
|
||||
/// Commits a drained batch of pending events in one transaction. All-or-nothing per
|
||||
/// batch: <c>site_events</c> ids are freshly minted GUIDs per <see cref="LogEventAsync"/>
|
||||
/// call (no cross-call replay/de-dup concern the way <c>SqliteAuditWriter</c> has), so
|
||||
/// unlike that reference there is no per-row duplicate-key swallow here — any failure
|
||||
/// rolls back the whole batch and faults every event in it. <c>site_events</c> is a
|
||||
/// CDC-replicated table (LocalDb Phase 1); its capture triggers fire per row on INSERT
|
||||
/// same as before, so batching the commit changes only how many rows land per
|
||||
/// transaction, not per-row trigger/replication behavior.
|
||||
/// </summary>
|
||||
private void FlushBatch(IReadOnlyList<PendingEvent> batch)
|
||||
{
|
||||
bool written;
|
||||
Exception? failure = null;
|
||||
|
||||
try
|
||||
{
|
||||
written = WithConnection(connection =>
|
||||
{
|
||||
using var transaction = connection.BeginTransaction();
|
||||
try
|
||||
{
|
||||
using var cmd = connection.CreateCommand();
|
||||
cmd.Transaction = transaction;
|
||||
cmd.CommandText = """
|
||||
INSERT INTO site_events (id, timestamp, event_type, severity, instance_id, source, message, details)
|
||||
VALUES ($id, $timestamp, $event_type, $severity, $instance_id, $source, $message, $details)
|
||||
""";
|
||||
cmd.Parameters.AddWithValue("$id", pending.Id);
|
||||
cmd.Parameters.AddWithValue("$timestamp", pending.Timestamp);
|
||||
cmd.Parameters.AddWithValue("$event_type", pending.EventType);
|
||||
cmd.Parameters.AddWithValue("$severity", pending.Severity);
|
||||
cmd.Parameters.AddWithValue("$instance_id", (object?)pending.InstanceId ?? DBNull.Value);
|
||||
cmd.Parameters.AddWithValue("$source", pending.Source);
|
||||
cmd.Parameters.AddWithValue("$message", pending.Message);
|
||||
cmd.Parameters.AddWithValue("$details", (object?)pending.Details ?? DBNull.Value);
|
||||
cmd.ExecuteNonQuery();
|
||||
});
|
||||
var pId = cmd.Parameters.Add("$id", SqliteType.Text);
|
||||
var pTimestamp = cmd.Parameters.Add("$timestamp", SqliteType.Text);
|
||||
var pEventType = cmd.Parameters.Add("$event_type", SqliteType.Text);
|
||||
var pSeverity = cmd.Parameters.Add("$severity", SqliteType.Text);
|
||||
var pInstanceId = cmd.Parameters.Add("$instance_id", SqliteType.Text);
|
||||
var pSource = cmd.Parameters.Add("$source", SqliteType.Text);
|
||||
var pMessage = cmd.Parameters.Add("$message", SqliteType.Text);
|
||||
var pDetails = cmd.Parameters.Add("$details", SqliteType.Text);
|
||||
|
||||
if (written)
|
||||
{
|
||||
pending.Completion.TrySetResult();
|
||||
foreach (var pending in batch)
|
||||
{
|
||||
pId.Value = pending.Id;
|
||||
pTimestamp.Value = pending.Timestamp;
|
||||
pEventType.Value = pending.EventType;
|
||||
pSeverity.Value = pending.Severity;
|
||||
pInstanceId.Value = (object?)pending.InstanceId ?? DBNull.Value;
|
||||
pSource.Value = pending.Source;
|
||||
pMessage.Value = pending.Message;
|
||||
pDetails.Value = (object?)pending.Details ?? DBNull.Value;
|
||||
cmd.ExecuteNonQuery();
|
||||
}
|
||||
|
||||
transaction.Commit();
|
||||
Interlocked.Increment(ref _committedBatchCount);
|
||||
}
|
||||
else
|
||||
catch
|
||||
{
|
||||
// WithConnection returns false only when the logger has been
|
||||
// disposed mid-drain; the event was not persisted. Fault the
|
||||
// Task instead of reporting false
|
||||
// success for a dropped audit event.
|
||||
pending.Completion.TrySetException(
|
||||
new ObjectDisposedException(nameof(SiteEventLogger),
|
||||
"Event could not be recorded: the event logger was disposed before the write completed."));
|
||||
transaction.Rollback();
|
||||
throw;
|
||||
}
|
||||
}
|
||||
catch (Exception ex)
|
||||
});
|
||||
}
|
||||
catch (Exception ex)
|
||||
{
|
||||
failure = ex;
|
||||
written = false;
|
||||
}
|
||||
|
||||
if (failure is not null)
|
||||
{
|
||||
// A write failure must be observable. Count every event in the failed
|
||||
// batch (Health Monitoring reads FailedWriteCount) and fault each
|
||||
// caller's Task instead of silently discarding the exception.
|
||||
Interlocked.Add(ref _failedWriteCount, batch.Count);
|
||||
_logger.LogError(failure, "Failed to record {Count} event(s) in batch (sqlite {SqliteError})",
|
||||
batch.Count, DescribeSqliteError(failure));
|
||||
foreach (var pending in batch)
|
||||
{
|
||||
// A write failure must be observable. Count it
|
||||
// (Health Monitoring reads FailedWriteCount) and fault the caller's
|
||||
// Task instead of silently discarding the exception.
|
||||
Interlocked.Increment(ref _failedWriteCount);
|
||||
_logger.LogError(ex, "Failed to record event: {EventType} from {Source} (sqlite {SqliteError})",
|
||||
pending.EventType, pending.Source, DescribeSqliteError(ex));
|
||||
pending.Completion.TrySetException(ex);
|
||||
pending.Completion.TrySetException(failure);
|
||||
}
|
||||
return;
|
||||
}
|
||||
|
||||
if (written)
|
||||
{
|
||||
foreach (var pending in batch)
|
||||
{
|
||||
pending.Completion.TrySetResult();
|
||||
}
|
||||
}
|
||||
else
|
||||
{
|
||||
// WithConnection returns false only when the logger has been
|
||||
// disposed mid-drain; none of the batch was persisted. Fault every
|
||||
// Task instead of reporting false success for a dropped event.
|
||||
var disposedEx = new ObjectDisposedException(nameof(SiteEventLogger),
|
||||
"Event could not be recorded: the event logger was disposed before the write completed.");
|
||||
foreach (var pending in batch)
|
||||
{
|
||||
pending.Completion.TrySetException(disposedEx);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user