Reduce App Insights exception noise

Author: Mateus Medeiros

.codex-skills/writefluency-deploy/SKILL.md

@@ -0,0 +1,87 @@
+---
+name: writefluency-deploy
+description: WriteFluency production deployment workflow for committing staged changes, pushing master, creating service version tags, and pushing those tags to trigger GitHub Actions deploys. Use when the user asks to deploy, push to production/prd, trigger production deploys, release staged changes, or create/push deploy tags for WriteFluency services.
+---
+
+# WriteFluency Deploy
+
+## Workflow
+
+Use this workflow for production deploys from `/Users/mateusmedeiros/Desktop/repos/WriteFluency`.
+
+1. Inspect the staged set and branch:
+
+```bash
+git status --short --branch
+git diff --cached --stat
+git diff --cached --name-only
+```
+
+Commit only staged changes unless the user explicitly asks to stage more. If there are no staged changes, stop and tell the user.
+
+2. Identify deploy targets from staged paths:
+
+- `src/propositions-service/**` or `tests/propositions-service/**` -> `propositions-v*`
+- `src/webapp/**` -> `webapp-v*`
+- `src/users-service/**` or `tests/users-service/**` -> `users-v*`
+- `src/users-progress-service/**`, `tests/users-progress-service/**`, or `infra/users-progress/**` -> check workflows before tagging; current workflows use `users-v*` for `deploy-users-progress.yml`, despite historical `users-progress-v*` tags.
+- `.github/workflows/deploy-*.yml`, `src/host/WriteFluency.AppHost/deploy-k8s.sh`, or `src/host/WriteFluency.AppHost/aspirate-overlays-*` -> inspect workflows and ask if service tags are not obvious.
+
+3. Verify tag triggers before creating tags:
+
+```bash
+rg -n "tags:|propositions-v|webapp-v|users-v|users-progress-v" .github/workflows
+git tag --sort=-creatordate | head -n 30
+```
+
+Choose the next patch version for each affected service. Examples:
+
+- latest `propositions-v1.0.8` -> next `propositions-v1.0.9`
+- latest `webapp-v1.0.9` -> next `webapp-v1.0.10`
+- latest `users-v1.1.16` -> next `users-v1.1.17`
+
+4. Commit staged changes with a concise message:
+
+```bash
+git commit -m "<message>"
+```
+
+5. Push `master` before tags:
+
+```bash
+git push origin master
+```
+
+6. Create tags on the pushed commit and push them:
+
+```bash
+git tag <service-vX.Y.Z> HEAD
+git push origin <service-vX.Y.Z> [<another-service-vX.Y.Z>...]
+```
+
+7. Confirm final state:
+
+```bash
+git status --short --branch
+git log --oneline -n 3 --decorate
+git ls-remote --tags origin <tag-1> <tag-2>
+```
+
+## Guardrails
+
+- Do not rewrite, delete, or move existing tags.
+- Do not run `git reset`, `git checkout --`, or otherwise discard work.
+- Do not commit unstaged changes unless the user explicitly asks.
+- If staged changes span multiple services, create one tag per affected service after pushing the commit.
+- If the branch is not `master`, ask before proceeding unless the user explicitly requested the current branch.
+- If `origin/master` has new commits, pull/rebase only after telling the user what will happen.
+- If tests were not already run in the session, run focused tests when the staged changes include code, unless the user asked for a fast deploy.
+
+## Known Tags
+
+As of the current repo workflow, production deploys are triggered by:
+
+- `.github/workflows/deploy-propositions.yml`: `propositions-v*`
+- `.github/workflows/deploy-webapp.yml`: `webapp-v*`
+- `.github/workflows/deploy-users.yml`: `users-v*`
+- `.github/workflows/deploy-users-progress.yml`: `users-v*`

.codex-skills/writefluency-deploy/agents/openai.yaml

@@ -0,0 +1,7 @@
+interface:
+  display_name: "WriteFluency Deploy"
+  short_description: "Commit and tag WriteFluency deploys."
+  default_prompt: "Use $writefluency-deploy to commit staged WriteFluency changes, push master, and create the correct production deploy tags."
+
+policy:
+  allow_implicit_invocation: true

infra/users-progress/main.bicep

@@ -83,6 +83,18 @@ param cosmosAccountName string
 @description('Cosmos DB endpoint URL.')
 param cosmosEndpoint string
 
+@description('Cosmos DB preferred regions for the East US Function App, in priority order.')
+param eastUsCosmosPreferredRegions array = [
+  'East US'
+  'Southeast Asia'
+]
+
+@description('Cosmos DB preferred regions for the Southeast Asia Function App, in priority order.')
+param southeastAsiaCosmosPreferredRegions array = [
+  'Southeast Asia'
+  'East US'
+]
+
 @description('Cosmos database name.')
 param cosmosDatabaseName string = 'wf-users-progress'
 
@@ -399,6 +411,16 @@ var corsAppSettings = [for (origin, index) in corsAllowedOrigins: {
   value: origin
 }]
 
+var eastUsCosmosPreferredRegionAppSettings = [for (region, index) in eastUsCosmosPreferredRegions: {
+  name: 'Cosmos__PreferredRegions__${index}'
+  value: region
+}]
+
+var southeastAsiaCosmosPreferredRegionAppSettings = [for (region, index) in southeastAsiaCosmosPreferredRegions: {
+  name: 'Cosmos__PreferredRegions__${index}'
+  value: region
+}]
+
 var telemetryTuningAppSettings = [
   {
     name: 'AzureFunctionsJobHost__logging__logLevel__default'
@@ -493,14 +515,18 @@ resource functionAppEastUs 'Microsoft.Web/sites@2024-04-01' = {
         maximumInstanceCount: maximumInstanceCount
         instanceMemoryMB: instanceMemoryMB
       }
+      #disable-next-line BCP037
+      siteUpdateStrategy: {
+        type: 'RollingUpdate'
+      }
     }
     siteConfig: {
       vnetRouteAllEnabled: true
       cors: {
         allowedOrigins: corsAllowedOrigins
         supportCredentials: true
       }
-      appSettings: concat(appSettingsBase, corsAppSettings, telemetryTuningAppSettings, [
+      appSettings: concat(appSettingsBase, eastUsCosmosPreferredRegionAppSettings, corsAppSettings, telemetryTuningAppSettings, [
         {
           name: 'AzureWebJobsStorage__accountName'
           value: storageEastUs.name
@@ -553,14 +579,18 @@ resource functionAppSoutheastAsia 'Microsoft.Web/sites@2024-04-01' = {
         maximumInstanceCount: maximumInstanceCount
         instanceMemoryMB: instanceMemoryMB
       }
+      #disable-next-line BCP037
+      siteUpdateStrategy: {
+        type: 'RollingUpdate'
+      }
     }
     siteConfig: {
       vnetRouteAllEnabled: true
       cors: {
         allowedOrigins: corsAllowedOrigins
         supportCredentials: true
       }
-      appSettings: concat(appSettingsBase, corsAppSettings, telemetryTuningAppSettings, [
+      appSettings: concat(appSettingsBase, southeastAsiaCosmosPreferredRegionAppSettings, corsAppSettings, telemetryTuningAppSettings, [
         {
           name: 'AzureWebJobsStorage__accountName'
           value: storageSoutheastAsia.name

src/host/WriteFluency.AppHost/aspirate-overlays-propositions/wf-api/deployment-patch.yaml

@@ -3,6 +3,7 @@ kind: Deployment
 metadata:
   name: wf-propositions-api
 spec:
+  replicas: 2
   strategy:
     type: RollingUpdate
     rollingUpdate:
@@ -14,6 +15,22 @@ spec:
       - name: ghcr-secret
       containers:
       - name: wf-propositions-api
+        readinessProbe:
+          httpGet:
+            path: /health
+            port: 8080
+          initialDelaySeconds: 10
+          periodSeconds: 10
+          timeoutSeconds: 3
+          failureThreshold: 6
+        livenessProbe:
+          httpGet:
+            path: /alive
+            port: 8080
+          initialDelaySeconds: 30
+          periodSeconds: 30
+          timeoutSeconds: 3
+          failureThreshold: 3
         envFrom:
         - configMapRef:
             name: wf-propositions-api-env

src/host/WriteFluency.AppHost/aspirate-overlays-users/wf-users-api/deployment-patch.yaml

@@ -3,6 +3,7 @@ kind: Deployment
 metadata:
   name: wf-users-api
 spec:
+  replicas: 2
   strategy:
     type: RollingUpdate
     rollingUpdate:
@@ -14,6 +15,22 @@ spec:
       - name: ghcr-secret
       containers:
       - name: wf-users-api
+        readinessProbe:
+          httpGet:
+            path: /health
+            port: 8080
+          initialDelaySeconds: 10
+          periodSeconds: 10
+          timeoutSeconds: 3
+          failureThreshold: 6
+        livenessProbe:
+          httpGet:
+            path: /alive
+            port: 8080
+          initialDelaySeconds: 30
+          periodSeconds: 30
+          timeoutSeconds: 3
+          failureThreshold: 3
         envFrom:
         - configMapRef:
             name: wf-users-api-env

src/propositions-service/WriteFluency.Infrastructure/Http/Services/ArticleExtractor.cs

@@ -65,6 +65,15 @@ public async Task<Result<string>> GetVisibleTextAsync(string url, CancellationTo
 
             return result;
         }
+        catch (HttpRequestException ex)
+        {
+            _logger.LogWarning(
+                "Article content was rejected or unavailable. Url={Url}, StatusCode={StatusCode}.",
+                url,
+                ex.StatusCode);
+
+            return Result.Fail(new Error("Failed to extract visible text").CausedBy(ex));
+        }
         catch (Exception ex)
         {
             _logger.LogError(ex, "Failed to extract visible text from {Url}", url);

src/users-progress-service/WriteFluency.UsersProgressService/Functions/UserProgressFunctions.cs

@@ -72,7 +72,15 @@ public async Task<HttpResponseData> StartAsync(
             return await CreateJsonAsync(request, HttpStatusCode.BadRequest, new { Error = "exercise_id_invalid" }, cancellationToken);
         }
 
-        var response = await _progressTrackingService.StartAsync(auth.UserId, body, cancellationToken);
+        ProgressOperationResponse response;
+        try
+        {
+            response = await _progressTrackingService.StartAsync(auth.UserId, body, cancellationToken);
+        }
+        catch (OperationCanceledException) when (cancellationToken.IsCancellationRequested)
+        {
+            return CreateClientClosedRequestResponse(request, "ProgressStart", auth.UserId);
+        }
 
         _logger.LogInformation(
             "Progress start request processed. UserId={UserId}, ExerciseId={ExerciseId}, Status={Status}, TrackingEnabled={TrackingEnabled}.",
@@ -128,7 +136,15 @@ public async Task<HttpResponseData> CompleteAsync(
             return await CreateJsonAsync(request, HttpStatusCode.BadRequest, new { Error = "exercise_id_invalid" }, cancellationToken);
         }
 
-        var response = await _progressTrackingService.CompleteAsync(auth.UserId, body, cancellationToken);
+        ProgressOperationResponse response;
+        try
+        {
+            response = await _progressTrackingService.CompleteAsync(auth.UserId, body, cancellationToken);
+        }
+        catch (OperationCanceledException) when (cancellationToken.IsCancellationRequested)
+        {
+            return CreateClientClosedRequestResponse(request, "ProgressComplete", auth.UserId);
+        }
 
         _logger.LogInformation(
             "Progress complete request processed. UserId={UserId}, ExerciseId={ExerciseId}, Status={Status}, TrackingEnabled={TrackingEnabled}.",
@@ -161,7 +177,15 @@ public async Task<HttpResponseData> SummaryAsync(
             return await CreateJsonAsync(request, HttpStatusCode.Unauthorized, new { Error = "unauthorized" }, cancellationToken);
         }
 
-        var response = await _progressTrackingService.GetSummaryAsync(auth.UserId, cancellationToken);
+        ProgressSummaryResponse response;
+        try
+        {
+            response = await _progressTrackingService.GetSummaryAsync(auth.UserId, cancellationToken);
+        }
+        catch (OperationCanceledException) when (cancellationToken.IsCancellationRequested)
+        {
+            return CreateClientClosedRequestResponse(request, "ProgressSummary", auth.UserId);
+        }
 
         _logger.LogInformation(
             "Progress summary request processed. UserId={UserId}, TotalItems={TotalItems}, InProgressCount={InProgressCount}, CompletedCount={CompletedCount}, TotalAttempts={TotalAttempts}.",
@@ -208,7 +232,15 @@ public async Task<HttpResponseData> StateAsync(
             return await CreateJsonAsync(request, HttpStatusCode.BadRequest, new { Error = "exercise_id_invalid" }, cancellationToken);
         }
 
-        var response = await _progressTrackingService.GetStateAsync(auth.UserId, exerciseId, cancellationToken);
+        ProgressStateResponse response;
+        try
+        {
+            response = await _progressTrackingService.GetStateAsync(auth.UserId, exerciseId, cancellationToken);
+        }
+        catch (OperationCanceledException) when (cancellationToken.IsCancellationRequested)
+        {
+            return CreateClientClosedRequestResponse(request, "ProgressState", auth.UserId);
+        }
 
         _logger.LogInformation(
             "Progress state request processed. UserId={UserId}, ExerciseId={ExerciseId}, HasServerState={HasServerState}, TrackingEnabled={TrackingEnabled}.",
@@ -241,7 +273,15 @@ public async Task<HttpResponseData> ItemsAsync(
             return await CreateJsonAsync(request, HttpStatusCode.Unauthorized, new { Error = "unauthorized" }, cancellationToken);
         }
 
-        var response = await _progressTrackingService.GetItemsAsync(auth.UserId, cancellationToken);
+        IReadOnlyList<ProgressItemResponse> response;
+        try
+        {
+            response = await _progressTrackingService.GetItemsAsync(auth.UserId, cancellationToken);
+        }
+        catch (OperationCanceledException) when (cancellationToken.IsCancellationRequested)
+        {
+            return CreateClientClosedRequestResponse(request, "ProgressItems", auth.UserId);
+        }
 
         _logger.LogInformation(
             "Progress items request processed. UserId={UserId}, ItemCount={ItemCount}.",
@@ -280,7 +320,15 @@ public async Task<HttpResponseData> AttemptsAsync(
             exerciseId = parsedExerciseId;
         }
 
-        var response = await _progressTrackingService.GetAttemptsAsync(auth.UserId, exerciseId, cancellationToken);
+        IReadOnlyList<ProgressAttemptResponse> response;
+        try
+        {
+            response = await _progressTrackingService.GetAttemptsAsync(auth.UserId, exerciseId, cancellationToken);
+        }
+        catch (OperationCanceledException) when (cancellationToken.IsCancellationRequested)
+        {
+            return CreateClientClosedRequestResponse(request, "ProgressAttempts", auth.UserId);
+        }
 
         _logger.LogInformation(
             "Progress attempts request processed. UserId={UserId}, ExerciseId={ExerciseId}, AttemptCount={AttemptCount}.",
@@ -335,7 +383,15 @@ public async Task<HttpResponseData> SaveStateAsync(
             return await CreateJsonAsync(request, HttpStatusCode.BadRequest, new { Error = "exercise_id_invalid" }, cancellationToken);
         }
 
-        var response = await _progressTrackingService.SaveStateAsync(auth.UserId, body, cancellationToken);
+        ProgressOperationResponse response;
+        try
+        {
+            response = await _progressTrackingService.SaveStateAsync(auth.UserId, body, cancellationToken);
+        }
+        catch (OperationCanceledException) when (cancellationToken.IsCancellationRequested)
+        {
+            return CreateClientClosedRequestResponse(request, "ProgressStateSave", auth.UserId);
+        }
 
         _logger.LogInformation(
             "Progress save-state request processed. UserId={UserId}, ExerciseId={ExerciseId}, Status={Status}, TrackingEnabled={TrackingEnabled}.",
@@ -409,6 +465,20 @@ private static bool IsAllowedOrigin(string uriLikeHeaderValue, HashSet<string> a
         return values.FirstOrDefault();
     }
 
+    private HttpResponseData CreateClientClosedRequestResponse(
+        HttpRequestData request,
+        string operationName,
+        string userId)
+    {
+        _logger.LogInformation(
+            "Progress request was canceled by the client. Operation={OperationName}, UserId={UserId}, Path={Path}.",
+            operationName,
+            userId,
+            request.Url.AbsolutePath);
+
+        return request.CreateResponse((HttpStatusCode)499);
+    }
+
     private static async Task<HttpResponseData> CreateJsonAsync(
         HttpRequestData request,
         HttpStatusCode statusCode,