Simple Chat

Operate SimpleChat day to day

Monitor, scale, review safety and feedback, manage data, and troubleshoot a running SimpleChat deployment.

What this covers

This guide gives administrators a day-to-day operating loop for SimpleChat after deployment and first configuration: watch telemetry, scale proven bottlenecks, review feedback and safety events, manage retention and backups, and know where to look when something breaks.

Video planned

Operate SimpleChat day to day walkthrough

Recording planned. Walk the full journey end to end and explain the decisions an admin makes along the way.

To publish this video:

  1. Upload the recording to YouTube or Microsoft Stream.
  2. Save a poster frame as docs/images/video-posters/guide-admin-operate-simplechat.png.
  3. Add the watch URL to this slot in docs/_data/media.yml.

Why it matters

SimpleChat depends on several Azure services that fail or saturate differently. App Service pressure, Cosmos RU exhaustion, Azure AI Search latency, model quota, extraction failures, and governance blocks need different responses. A steady operating rhythm keeps admins from guessing during incidents.

Before you start

  • Confirm Logging settings are configured for Application Insights and that temporary debug logging has an auto-turnoff window when used.
  • Confirm Backup, Migrate & Restore settings have a valid storage target and a retention policy.
  • Know which optional capabilities are enabled so you can check their backing Azure resources during incidents.
  • Keep access to Azure Portal views for App Service, Cosmos DB, Azure AI Search, Azure OpenAI or Foundry, Document Intelligence, Storage, Redis, and Application Insights.

Step 1: Monitor the application before changing settings

Use Application Insights and App Service logs to establish whether the issue is in the web tier, instrumentation, configuration, or a downstream dependency. The troubleshooting reference starts with failed requests and exceptions because those records identify the failing operation before you change Admin Settings.

Screenshot needed

Monitor requests and exceptions

Capture Application Insights requests and exceptions with operation identifiers visible and user content redacted.

Save as docs/images/guides/admin-operate-simplechat-monitor.png

Drop the file at that path and it replaces this card automatically. No other change needed.

Step 2: Scale the bottleneck you can prove

Use Application Scaling and Scale settings together. Scale App Service up for memory or CPU pressure and out for concurrent traffic. Watch Cosmos RU consumption and HTTP 429 responses before increasing throughput. Use Azure AI Search replicas, partitions, or higher tiers when query latency, indexing throughput, or service limits justify it. Treat AI service quota and rate limits as separate from web-tier capacity.

Step 3: Review safety violations, feedback, and governed items

Use Safety settings to control Content Safety, feedback collection, review role requirements, and conversation archiving. Use Governance settings to review personal, group, and global endpoints, agents, actions, and MCP destinations. Keep review roles assigned before enforcing review-only access.

Screenshot needed

Review operational queues

Capture safety, feedback, or governance review queues with any user content blurred or replaced by sample data.

Save as docs/images/guides/admin-operate-simplechat-review.png

Drop the file at that path and it replaces this card automatically. No other change needed.

Step 4: Manage retention, archiving, and backups

Use Workspaces settings for default document and conversation retention by workspace type. Use Safety settings when deletion should archive conversations instead of removing them immediately. Use Backup, Migrate & Restore settings for scheduled backups, restore preflight, job history, migration, and emergency Cosmos JSON edits.

Step 5: Keep ingestion and source data under control

Use Search and Extract settings to tune extraction, chunking, URL Access, Web Search, Deep Research, Speech, and Video Indexer. Use File Sync settings to limit source counts, schedule frequency, files per run, GB per run, concurrency, recursive sources, and allowed workspace scopes before sync jobs grow unexpectedly.

Step 6: Diagnose failures by layer

Start with Troubleshooting. Query failed requests, pivot to exceptions with the same operation_Id, and only then decide whether the problem is authentication, model routing, retrieval, extraction, storage, cache, or a feature toggle. Use temporary debug logging for short windows and turn it off after reproducing the issue.

Screenshot needed

Diagnose by layer

Capture the troubleshooting flow or a monitoring workbook that separates app, data, search, and AI service checks.

Save as docs/images/guides/admin-operate-simplechat-diagnose.png

Drop the file at that path and it replaces this card automatically. No other change needed.

Verify it worked

  • Routine telemetry shows successful requests, expected latency, and no unexplained exception spike.
  • Backup jobs complete on schedule and old backups follow the configured retention policy.
  • Safety, feedback, governance, and file-processing queues are reviewed at the cadence your tenant requires.
  • Scaling changes are tied to observed pressure such as CPU, memory, RU saturation, search latency, or AI service quota.
  • Temporary debug logging is off outside active investigations.

Troubleshooting

Symptom Likely cause Fix
You cannot tell which backend call failed The investigation started from settings instead of telemetry Query Application Insights failed requests, capture operation_Id, then pivot to exceptions.
Horizontal scale causes inconsistent behavior Shared cache or session assumptions were not ready Review Redis-backed cache guidance in Scale settings before serious scale-out.
File processing history grows without cleanup File-processing logs are enabled without a retention routine Use Logging settings cleanup controls and set an operating cadence.