Experiment with ES|QL lookup joins for event and stack queries

.github/workflows/build.yaml

@@ -134,7 +134,6 @@ jobs:
       - name: Resolve Elasticsearch image
         id: elasticsearch_image
         env:
-          PR_BASE_SHA: ${{ github.event.pull_request.base.sha }}
           PR_HEAD_REPOSITORY: ${{ github.event.pull_request.head.repo.full_name }}
           PREVIEW_BASE_SHA: ${{ github.event.client_payload.base_sha }}
           PUSH_BEFORE_SHA: ${{ github.event.before }}
@@ -145,7 +144,7 @@ jobs:
           build_locally=false
 
           if [[ "$GITHUB_EVENT_NAME" == "pull_request" ]] &&
-            ! git diff --quiet "$PR_BASE_SHA"...HEAD -- build/docker/elasticsearch/9.x .github/workflows/elasticsearch-docker-9.yml; then
+            ! git diff --quiet origin/main...HEAD -- build/docker/elasticsearch/9.x .github/workflows/elasticsearch-docker-9.yml; then
             image_changed=true
           elif [[ "$GITHUB_EVENT_NAME" == "push" && "$GITHUB_REF" == "refs/heads/main" ]] &&
             ! git diff --quiet "$PUSH_BEFORE_SHA"..HEAD -- build/docker/elasticsearch/9.x .github/workflows/elasticsearch-docker-9.yml; then
@@ -362,8 +361,8 @@ jobs:
       - name: Wait for Aspire Resources
         run: |
           for attempt in {1..60}; do
-            if curl -fksS https://web-ex.dev.localhost:7131/api/v2/about > /dev/null &&
-              curl -fksS https://web-ex.dev.localhost:7131/next/login > /dev/null; then
+            if curl -fksS --connect-timeout 5 --max-time 10 https://web-ex.dev.localhost:7131/api/v2/about > /dev/null &&
+              curl -fksS --connect-timeout 5 --max-time 10 https://web-ex.dev.localhost:7131/next/login > /dev/null; then
               break
             fi
 
@@ -382,8 +381,8 @@ jobs:
 
       - name: Verify E2E Endpoints
         run: |
-          curl -fksS https://web-ex.dev.localhost:7131/api/v2/about > /dev/null
-          curl -fksS https://web-ex.dev.localhost:7131/next/login > /dev/null
+          curl -fksS --connect-timeout 5 --max-time 10 https://web-ex.dev.localhost:7131/api/v2/about > /dev/null
+          curl -fksS --connect-timeout 5 --max-time 10 https://web-ex.dev.localhost:7131/next/login > /dev/null
 
       - name: Run Playwright E2E Tests
         working-directory: src/Exceptionless.Web/ClientApp

src/Exceptionless.AppHost/Exceptionless.AppHost.csproj

@@ -12,7 +12,6 @@
     <PackageReference Include="Aspire.Hosting.Browsers" Version="13.4.6-preview.1.26319.6" />
     <PackageReference Include="Aspire.Hosting.JavaScript" Version="13.4.6" />
     <PackageReference Include="Aspire.Hosting.Redis" Version="13.4.6" />
-    <PackageReference Include="AspNetCore.HealthChecks.Elasticsearch" Version="9.0.0" />
     <PackageReference Include="CommunityToolkit.Aspire.Hosting.Deno" Version="13.4.0" />
   </ItemGroup>
 

src/Exceptionless.AppHost/Extensions/ElasticsearchExtensions.cs

@@ -1,4 +1,4 @@
-using Elastic.Clients.Elasticsearch;
+using System.Text.Json;
 using Microsoft.Extensions.DependencyInjection;
 using Microsoft.Extensions.Diagnostics.HealthChecks;
 
@@ -136,19 +136,25 @@ public async Task<HealthCheckResult> CheckHealthAsync(HealthCheckContext context
         if (string.IsNullOrEmpty(connectionString))
             return new HealthCheckResult(context.Registration.FailureStatus, "Connection string not available.");
 
-        using var settings = new ElasticsearchClientSettings(new Uri(connectionString));
-        var client = new ElasticsearchClient(settings);
-        var response = await client.Cluster.HealthAsync(
-            request => request.WaitForStatus(Elastic.Clients.Elasticsearch.HealthStatus.Yellow),
-            cancellationToken);
-        bool isReady = response.IsValidResponse
-            && !response.TimedOut
-            && response.Status is Elastic.Clients.Elasticsearch.HealthStatus.Yellow or Elastic.Clients.Elasticsearch.HealthStatus.Green;
-        if (isReady)
-            return HealthCheckResult.Healthy();
-
-        return new HealthCheckResult(
-            context.Registration.FailureStatus,
-            $"Elasticsearch cluster health check failed. Timed out: {response.TimedOut}; status: {response.Status}. {response.DebugInformation}");
+        try
+        {
+            using var client = new HttpClient { BaseAddress = new Uri(connectionString), Timeout = TimeSpan.FromSeconds(10) };
+            using var response = await client.GetAsync("_cluster/health?wait_for_status=yellow&timeout=5s", cancellationToken);
+            if (!response.IsSuccessStatusCode)
+                return new HealthCheckResult(context.Registration.FailureStatus, $"Elasticsearch cluster health returned HTTP {(int)response.StatusCode}.");
+
+            await using var responseStream = await response.Content.ReadAsStreamAsync(cancellationToken);
+            using var document = await JsonDocument.ParseAsync(responseStream, cancellationToken: cancellationToken);
+            bool timedOut = document.RootElement.TryGetProperty("timed_out", out var timedOutElement) && timedOutElement.GetBoolean();
+            string? status = document.RootElement.TryGetProperty("status", out var statusElement) ? statusElement.GetString() : null;
+            if (!timedOut && status is "yellow" or "green")
+                return HealthCheckResult.Healthy();
+
+            return new HealthCheckResult(context.Registration.FailureStatus, $"Elasticsearch cluster health check failed. Timed out: {timedOut}; status: {status ?? "unknown"}.");
+        }
+        catch (Exception ex) when (ex is HttpRequestException or JsonException or TaskCanceledException)
+        {
+            return new HealthCheckResult(context.Registration.FailureStatus, "Elasticsearch cluster health check request failed.", ex);
+        }
     }
 }

src/Exceptionless.AppHost/appsettings.Development.json

@@ -1,4 +1,11 @@
 {
+  "Elasticsearch": {
+    "Port": 9215,
+    "ContainerName": "Exceptionless-Elasticsearch-9-5-Lookup",
+    "DataVolume": "exceptionless.elasticsearch-9-5-lookup.data.v1",
+    "KibanaPort": 5615,
+    "KibanaContainerName": "Exceptionless-Kibana-9-5-Lookup"
+  },
   "Logging": {
     "LogLevel": {
       "Default": "Information",

src/Exceptionless.Core/Bootstrapper.cs

@@ -77,6 +77,7 @@ public static void RegisterServices(IServiceCollection services, AppOptions appO
 
         services.AddSingleton<ExceptionlessElasticConfiguration>();
         services.AddSingleton<ElasticsearchClient>(s => s.GetRequiredService<ExceptionlessElasticConfiguration>().Client);
+        services.AddSingleton<IStackRollupSearchService, StackRollupSearchService>();
         services.AddSingleton<IElasticConfiguration>(s => s.GetRequiredService<ExceptionlessElasticConfiguration>());
         services.AddStartupAction<ExceptionlessElasticConfiguration>();
 

src/Exceptionless.Core/Repositories/Configuration/Indexes/StackIndex.cs

@@ -17,7 +17,7 @@ public sealed class StackIndex : VersionedIndex<Stack>
 
     private readonly ExceptionlessElasticConfiguration _configuration;
 
-    public StackIndex(ExceptionlessElasticConfiguration configuration) : base(configuration, configuration.Options.ScopePrefix + "stacks", 1)
+    public StackIndex(ExceptionlessElasticConfiguration configuration) : base(configuration, configuration.Options.ScopePrefix + "stacks", 2)
     {
         _configuration = configuration;
     }
@@ -27,11 +27,28 @@ public override void ConfigureIndex(CreateIndexRequestDescriptor idx)
         base.ConfigureIndex(idx);
         idx.Settings(s => s
             .Analysis(a => BuildAnalysis(a))
-            .NumberOfShards(_configuration.Options.NumberOfShards)
+            .Mode("lookup")
+            .NumberOfShards(1)
             .NumberOfReplicas(_configuration.Options.NumberOfReplicas)
             .Priority(5));
     }
 
+    protected override Task UpdateIndexAsync(string name, Action<PutIndicesSettingsRequestDescriptor>? descriptor = null)
+    {
+        if (descriptor is not null)
+            return base.UpdateIndexAsync(name, descriptor);
+
+        // index.mode is a final creation-only setting. Foundatio derives updates from
+        // ConfigureIndex, so keep the mutable settings explicit for existing indexes.
+        return base.UpdateIndexAsync(name, update => update
+            .Reopen(true)
+            .Settings(new IndexSettings
+            {
+                NumberOfReplicas = _configuration.Options.NumberOfReplicas,
+                Priority = 5
+            }));
+    }
+
     public override void ConfigureIndexMapping(TypeMappingDescriptor<Stack> map)
     {
         map

src/Exceptionless.Core/Repositories/Queries/Visitors/EventStackFilterQueryVisitor.cs

@@ -65,15 +65,15 @@ public EventStackFilter()
         _stackQueryVisitor.AddVisitor(new RemoveFieldsQueryVisitor(f => !stackFields.Contains(f)));
         _stackQueryVisitor.AddVisitor(new CleanupQueryVisitor());
         // handles stack special fields and changing event field names to their stack equivalent
-        _stackQueryVisitor.AddVisitor(new StackFilterQueryVisitor());
+        _stackQueryVisitor.AddVisitor(new StackFilterQueryVisitor(_stackOnlyFields.Union(_stackOnlySpecialFields)));
         _stackQueryVisitor.AddVisitor(new CleanupQueryVisitor());
 
         _invertedStackQueryVisitor = new ChainedQueryVisitor();
         // remove everything not in the stack fields list
         _invertedStackQueryVisitor.AddVisitor(new RemoveFieldsQueryVisitor(f => !stackFields.Contains(f)));
         _invertedStackQueryVisitor.AddVisitor(new CleanupQueryVisitor());
         // handles stack special fields and changing event field names to their stack equivalent
-        _invertedStackQueryVisitor.AddVisitor(new StackFilterQueryVisitor());
+        _invertedStackQueryVisitor.AddVisitor(new StackFilterQueryVisitor(_stackOnlyFields.Union(_stackOnlySpecialFields)));
         _invertedStackQueryVisitor.AddVisitor(new CleanupQueryVisitor());
         // inverts the filter
         _invertedStackQueryVisitor.AddVisitor(new InvertQueryVisitor(_stackNonInvertedFields));
@@ -109,13 +109,21 @@ public EventStackFilter()
             InvertedFilter = invertedResult?.ToString(),
             HasStatus = context.GetBoolean(nameof(StackFilter.HasStatus)),
             HasStackIds = context.GetBoolean(nameof(StackFilter.HasStackIds)),
-            HasStatusOpen = context.GetBoolean(nameof(StackFilter.HasStatusOpen))
+            HasStatusOpen = context.GetBoolean(nameof(StackFilter.HasStatusOpen)),
+            HasStackOnlyCriteria = context.GetBoolean(nameof(StackFilter.HasStackOnlyCriteria))
         };
     }
 }
 
 public class StackFilterQueryVisitor : ChainableQueryVisitor
 {
+    private readonly ISet<string> _stackOnlyFields;
+
+    public StackFilterQueryVisitor(IEnumerable<string>? stackOnlyFields = null)
+    {
+        _stackOnlyFields = new HashSet<string>(stackOnlyFields ?? [], StringComparer.OrdinalIgnoreCase);
+    }
+
     public override Task<IQueryNode?> VisitAsync(TermNode node, IQueryVisitorContext context)
     {
         IQueryNode result = node;
@@ -127,6 +135,9 @@ public class StackFilterQueryVisitor : ChainableQueryVisitor
             return Task.FromResult<IQueryNode?>(null);
         }
 
+        if (_stackOnlyFields.Contains(node.Field))
+            context.SetValue(nameof(StackFilter.HasStackOnlyCriteria), true);
+
         // process special stack fields
         switch (node.Field?.ToLowerInvariant())
         {
@@ -216,4 +227,5 @@ public record StackFilter
     public required bool HasStatus { get; set; }
     public required bool HasStatusOpen { get; set; }
     public required bool HasStackIds { get; set; }
+    public required bool HasStackOnlyCriteria { get; set; }
 }