Update Runner to send step updates to Results (#2510)

* Also send Steps update to Results service

* Refactor to separate results server from current job server

* If hit any error while uploading to Results, skip Results upload

* Add proxy authentication and buffer request for WinHttpHandler

* Remove unnecessary null guard

* Also send Results telemetry when step update fails

* IResultsServer is not disposable

src/Runner.Common/JobServer.cs

@@ -24,15 +24,11 @@ namespace GitHub.Runner.Common
         Task ConnectAsync(VssConnection jobConnection);
 
         void InitializeWebsocketClient(ServiceEndpoint serviceEndpoint);
-        void InitializeResultsClient(Uri uri, string token);
 
         // logging and console
         Task<TaskLog> AppendLogContentAsync(Guid scopeIdentifier, string hubName, Guid planId, int logId, Stream uploadStream, CancellationToken cancellationToken);
         Task AppendTimelineRecordFeedAsync(Guid scopeIdentifier, string hubName, Guid planId, Guid timelineId, Guid timelineRecordId, Guid stepId, IList<string> lines, long? startLine, CancellationToken cancellationToken);
         Task<TaskAttachment> CreateAttachmentAsync(Guid scopeIdentifier, string hubName, Guid planId, Guid timelineId, Guid timelineRecordId, String type, String name, Stream uploadStream, CancellationToken cancellationToken);
-        Task CreateStepSummaryAsync(string planId, string jobId, Guid stepId, string file, CancellationToken cancellationToken);
-        Task CreateResultsStepLogAsync(string planId, string jobId, Guid stepId, string file, bool finalize, bool firstBlock, long lineCount, CancellationToken cancellationToken);
-        Task CreateResultsJobLogAsync(string planId, string jobId, string file, bool finalize, bool firstBlock, long lineCount, CancellationToken cancellationToken);
         Task<TaskLog> CreateLogAsync(Guid scopeIdentifier, string hubName, Guid planId, TaskLog log, CancellationToken cancellationToken);
         Task<Timeline> CreateTimelineAsync(Guid scopeIdentifier, string hubName, Guid planId, Guid timelineId, CancellationToken cancellationToken);
         Task<List<TimelineRecord>> UpdateTimelineRecordsAsync(Guid scopeIdentifier, string hubName, Guid planId, Guid timelineId, IEnumerable<TimelineRecord> records, CancellationToken cancellationToken);
@@ -46,7 +42,6 @@ namespace GitHub.Runner.Common
         private bool _hasConnection;
         private VssConnection _connection;
         private TaskHttpClient _taskClient;
-        private ResultsHttpClient _resultsClient;
         private ClientWebSocket _websocketClient;
 
         private ServiceEndpoint _serviceEndpoint;
@@ -150,12 +145,6 @@ namespace GitHub.Runner.Common
             InitializeWebsocketClient(TimeSpan.Zero);
         }
 
-        public void InitializeResultsClient(Uri uri, string token)
-        {
-            var httpMessageHandler = HostContext.CreateHttpClientHandler();
-            this._resultsClient = new ResultsHttpClient(uri, httpMessageHandler, token, disposeHandler: true);
-        }
-
         public ValueTask DisposeAsync()
         {
             CloseWebSocket(WebSocketCloseStatus.NormalClosure, CancellationToken.None);
@@ -318,32 +307,6 @@ namespace GitHub.Runner.Common
             return _taskClient.CreateAttachmentAsync(scopeIdentifier, hubName, planId, timelineId, timelineRecordId, type, name, uploadStream, cancellationToken: cancellationToken);
         }
 
-        public Task CreateStepSummaryAsync(string planId, string jobId, Guid stepId, string file, CancellationToken cancellationToken)
-        {
-            if (_resultsClient != null)
-            {
-                return _resultsClient.UploadStepSummaryAsync(planId, jobId, stepId, file, cancellationToken: cancellationToken);
-            }
-            throw new InvalidOperationException("Results client is not initialized.");
-        }
-
-        public Task CreateResultsStepLogAsync(string planId, string jobId, Guid stepId, string file, bool finalize, bool firstBlock, long lineCount, CancellationToken cancellationToken)
-        {
-            if (_resultsClient != null)
-            {
-                return _resultsClient.UploadResultsStepLogAsync(planId, jobId, stepId, file, finalize, firstBlock, lineCount, cancellationToken: cancellationToken);
-            }
-            throw new InvalidOperationException("Results client is not initialized.");
-        }
-
-        public Task CreateResultsJobLogAsync(string planId, string jobId, string file, bool finalize, bool firstBlock, long lineCount, CancellationToken cancellationToken)
-        {
-            if (_resultsClient != null)
-            {
-                return _resultsClient.UploadResultsJobLogAsync(planId, jobId, file, finalize, firstBlock, lineCount, cancellationToken: cancellationToken);
-            }
-            throw new InvalidOperationException("Results client is not initialized.");
-        }
 
         public Task<TaskLog> CreateLogAsync(Guid scopeIdentifier, string hubName, Guid planId, TaskLog log, CancellationToken cancellationToken)
         {

src/Runner.Common/JobServerQueue.cs

@@ -65,6 +65,7 @@ namespace GitHub.Runner.Common
 
         // common
         private IJobServer _jobServer;
+        private IResultsServer _resultsServer;
         private Task[] _allDequeueTasks;
         private readonly TaskCompletionSource<int> _jobCompletionSource = new();
         private readonly TaskCompletionSource<int> _jobRecordUpdated = new();
@@ -91,6 +92,7 @@ namespace GitHub.Runner.Common
         {
             base.Initialize(hostContext);
             _jobServer = hostContext.GetService<IJobServer>();
+            _resultsServer = hostContext.GetService<IResultsServer>();
         }
 
         public void Start(Pipelines.AgentJobRequestMessage jobRequest)
@@ -111,7 +113,7 @@ namespace GitHub.Runner.Common
                 !string.IsNullOrEmpty(resultsReceiverEndpoint))
             {
                 Trace.Info("Initializing results client");
-                _jobServer.InitializeResultsClient(new Uri(resultsReceiverEndpoint), accessToken);
+                _resultsServer.InitializeResultsClient(new Uri(resultsReceiverEndpoint), accessToken);
                 _resultsClientInitiated = true;
             }
 
@@ -512,19 +514,14 @@ namespace GitHub.Runner.Common
                         }
                         catch (Exception ex)
                         {
-                            var issue = new Issue() { Type = IssueType.Warning, Message = $"Caught exception during file upload to results. {ex.Message}" };
-                            issue.Data[Constants.Runner.InternalTelemetryIssueDataKey] = Constants.Runner.ResultsUploadFailure;
-
-                            var telemetryRecord = new TimelineRecord()
-                            {
-                                Id = Constants.Runner.TelemetryRecordId,
-                            };
-                            telemetryRecord.Issues.Add(issue);
-                            QueueTimelineRecordUpdate(_jobTimelineId, telemetryRecord);
-
                             Trace.Info("Catch exception during file upload to results, keep going since the process is best effort.");
                             Trace.Error(ex);
                             errorCount++;
+
+                            // If we hit any exceptions uploading to Results, let's skip any additional uploads to Results
+                            _resultsClientInitiated = false;
+
+                            SendResultsTelemetry(ex);
                         }
                     }
 
@@ -542,6 +539,19 @@ namespace GitHub.Runner.Common
             }
         }
 
+        private void SendResultsTelemetry(Exception ex)
+        {
+            var issue = new Issue() { Type = IssueType.Warning, Message = $"Caught exception with results. {ex.Message}" };
+            issue.Data[Constants.Runner.InternalTelemetryIssueDataKey] = Constants.Runner.ResultsUploadFailure;
+
+            var telemetryRecord = new TimelineRecord()
+            {
+                Id = Constants.Runner.TelemetryRecordId,
+            };
+            telemetryRecord.Issues.Add(issue);
+            QueueTimelineRecordUpdate(_jobTimelineId, telemetryRecord);
+        }
+
         private async Task ProcessTimelinesUpdateQueueAsync(bool runOnce = false)
         {
             while (!_jobCompletionSource.Task.IsCompleted || runOnce)
@@ -612,6 +622,22 @@ namespace GitHub.Runner.Common
                         try
                         {
                             await _jobServer.UpdateTimelineRecordsAsync(_scopeIdentifier, _hubName, _planId, update.TimelineId, update.PendingRecords, default(CancellationToken));
+                            try
+                            {
+                                if (_resultsClientInitiated)
+                                {
+                                    await _resultsServer.UpdateResultsWorkflowStepsAsync(_scopeIdentifier, _hubName, _planId, update.TimelineId, update.PendingRecords, default(CancellationToken));
+                                }
+                            }
+                            catch (Exception e)
+                            {
+                                Trace.Info("Catch exception during update steps, skip update Results.");
+                                Trace.Error(e);
+                                _resultsClientInitiated = false;
+
+                                SendResultsTelemetry(e);
+                            }
+
                             if (_bufferedRetryRecords.Remove(update.TimelineId))
                             {
                                 Trace.Verbose("Cleanup buffered timeline record for timeline: {0}.", update.TimelineId);
@@ -819,7 +845,7 @@ namespace GitHub.Runner.Common
             Trace.Info($"Starting to upload summary file to results service {file.Name}, {file.Path}");
             ResultsFileUploadHandler summaryHandler = async (file) =>
             {
-                await _jobServer.CreateStepSummaryAsync(file.PlanId, file.JobId, file.RecordId, file.Path, CancellationToken.None);
+                await _resultsServer.CreateResultsStepSummaryAsync(file.PlanId, file.JobId, file.RecordId, file.Path, CancellationToken.None);
             };
 
             await UploadResultsFile(file, summaryHandler);
@@ -830,7 +856,7 @@ namespace GitHub.Runner.Common
             Trace.Info($"Starting upload of step log file to results service {file.Name}, {file.Path}");
             ResultsFileUploadHandler stepLogHandler = async (file) =>
             {
-                await _jobServer.CreateResultsStepLogAsync(file.PlanId, file.JobId, file.RecordId, file.Path, file.Finalize, file.FirstBlock, file.TotalLines, CancellationToken.None);
+                await _resultsServer.CreateResultsStepLogAsync(file.PlanId, file.JobId, file.RecordId, file.Path, file.Finalize, file.FirstBlock, file.TotalLines, CancellationToken.None);
             };
 
             await UploadResultsFile(file, stepLogHandler);
@@ -841,7 +867,7 @@ namespace GitHub.Runner.Common
             Trace.Info($"Starting upload of job log file to results service {file.Name}, {file.Path}");
             ResultsFileUploadHandler jobLogHandler = async (file) =>
             {
-                await _jobServer.CreateResultsJobLogAsync(file.PlanId, file.JobId, file.Path, file.Finalize, file.FirstBlock, file.TotalLines, CancellationToken.None);
+                await _resultsServer.CreateResultsJobLogAsync(file.PlanId, file.JobId, file.Path, file.Finalize, file.FirstBlock, file.TotalLines, CancellationToken.None);
             };
 
             await UploadResultsFile(file, jobLogHandler);
@@ -849,6 +875,11 @@ namespace GitHub.Runner.Common
 
         private async Task UploadResultsFile(ResultsUploadFileInfo file, ResultsFileUploadHandler uploadHandler)
         {
+            if (!_resultsClientInitiated)
+            {
+                return;
+            }
+
             bool uploadSucceed = false;
             try
             {
@@ -903,8 +934,6 @@ namespace GitHub.Runner.Common
         public long TotalLines { get; set; }
     }
 
-
-
     internal class ConsoleLineInfo
     {
         public ConsoleLineInfo(Guid recordId, string line, long? lineNumber)

src/Runner.Common/ResultsServer.cs

@@ -0,0 +1,98 @@
+using System;
+using System.Collections.Generic;
+using System.Linq;
+using System.Threading;
+using System.Threading.Tasks;
+using GitHub.DistributedTask.WebApi;
+using GitHub.Services.Results.Client;
+
+namespace GitHub.Runner.Common
+{
+    [ServiceLocator(Default = typeof(ResultServer))]
+    public interface IResultsServer : IRunnerService
+    {
+        void InitializeResultsClient(Uri uri, string token);
+
+        // logging and console
+        Task CreateResultsStepSummaryAsync(string planId, string jobId, Guid stepId, string file,
+            CancellationToken cancellationToken);
+
+        Task CreateResultsStepLogAsync(string planId, string jobId, Guid stepId, string file, bool finalize,
+            bool firstBlock, long lineCount, CancellationToken cancellationToken);
+
+        Task CreateResultsJobLogAsync(string planId, string jobId, string file, bool finalize, bool firstBlock,
+            long lineCount, CancellationToken cancellationToken);
+
+        Task UpdateResultsWorkflowStepsAsync(Guid scopeIdentifier, string hubName, Guid planId, Guid timelineId,
+            IEnumerable<TimelineRecord> records, CancellationToken cancellationToken);
+    }
+
+    public sealed class ResultServer : RunnerService, IResultsServer
+    {
+        private ResultsHttpClient _resultsClient;
+
+        public void InitializeResultsClient(Uri uri, string token)
+        {
+            var httpMessageHandler = HostContext.CreateHttpClientHandler();
+            this._resultsClient = new ResultsHttpClient(uri, httpMessageHandler, token, disposeHandler: true);
+        }
+
+        public Task CreateResultsStepSummaryAsync(string planId, string jobId, Guid stepId, string file,
+            CancellationToken cancellationToken)
+        {
+            if (_resultsClient != null)
+            {
+                return _resultsClient.UploadStepSummaryAsync(planId, jobId, stepId, file,
+                    cancellationToken: cancellationToken);
+            }
+
+            throw new InvalidOperationException("Results client is not initialized.");
+        }
+
+        public Task CreateResultsStepLogAsync(string planId, string jobId, Guid stepId, string file, bool finalize,
+            bool firstBlock, long lineCount, CancellationToken cancellationToken)
+        {
+            if (_resultsClient != null)
+            {
+                return _resultsClient.UploadResultsStepLogAsync(planId, jobId, stepId, file, finalize, firstBlock,
+                    lineCount, cancellationToken: cancellationToken);
+            }
+
+            throw new InvalidOperationException("Results client is not initialized.");
+        }
+
+        public Task CreateResultsJobLogAsync(string planId, string jobId, string file, bool finalize, bool firstBlock,
+            long lineCount, CancellationToken cancellationToken)
+        {
+            if (_resultsClient != null)
+            {
+                return _resultsClient.UploadResultsJobLogAsync(planId, jobId, file, finalize, firstBlock, lineCount,
+                    cancellationToken: cancellationToken);
+            }
+
+            throw new InvalidOperationException("Results client is not initialized.");
+        }
+
+        public Task UpdateResultsWorkflowStepsAsync(Guid scopeIdentifier, string hubName, Guid planId, Guid timelineId,
+            IEnumerable<TimelineRecord> records, CancellationToken cancellationToken)
+        {
+            if (_resultsClient != null)
+            {
+                try
+                {
+                    var timelineRecords = records.ToList();
+                    return _resultsClient.UpdateWorkflowStepsAsync(planId, new List<TimelineRecord>(timelineRecords),
+                        cancellationToken: cancellationToken);
+                }
+                catch (Exception ex)
+                {
+                    // Log error, but continue as this call is best-effort
+                    Trace.Info($"Failed to update steps status due to {ex.GetType().Name}");
+                    Trace.Error(ex);
+                }
+            }
+
+            throw new InvalidOperationException("Results client is not initialized.");
+        }
+    }
+}