OBSERVABILITY
Findable runs in your Azure tenant, so it reports to your tooling. Application logs stream into your own Log Analytics workspace, every conversation is audited in your own Cosmos DB, and health checks surface in an admin console your team already has access to. Nothing routes through a vendor dashboard you can't query.
Findable ships application logs directly into an Azure Log Analytics workspace you own, over the Monitor Logs Ingestion API. Authentication is managed identity against a Data Collection Rule in your subscription — no API keys to rotate, no shared secret, no vendor-side log store. You point Findable at a Data Collection Endpoint, grant the app's identity the Monitoring Metrics Publisher role on your rule, and logs start flowing within minutes. No restart required.
Because the logs are ordinary rows in your workspace, they're ordinary KQL. Chart error rates over time, break volume down by log level, extract latency numbers out of structured metadata and graph the trend — then pin any of it to an Azure dashboard or wire it to an alert rule. Your retention policy, your access controls, your cost plan. Findable doesn't own the data and can't hold it hostage.
Log shipping is non-blocking by construction. Records buffer in memory and upload out-of-band in batches; the buffer is bounded, so a slow endpoint drops old rows rather than growing without limit. Upload failures are isolated — a bad workspace, a missing role assignment, or a network partition degrades your dashboards and nothing else. A final flush runs on shutdown, and a built-in health probe tells you whether the identity can actually reach Azure Monitor before you go looking in the wrong place.
CONVERSATION AUDIT
Prompts, responses, the model that answered, the sources it cited, and the user who asked are written to a chat log container in your own Cosmos DB. Admins can review, filter, and export; retention runs on a schedule you set.
Thumbs-up and thumbs-down feedback is captured alongside the conversation, so the quality signal and the transcript live together instead of in a vendor's analytics product.
SYSTEM HEALTH
A connectivity console probes every dependency Findable touches — model endpoints, search services, blob storage, database connections, MCP servers, notification providers, and the Azure Monitor pipeline — and reports each one independently.
Scheduled jobs for permission resync and retention sweeps report their last run, status, and message on the same page, with a run-now button when you don't want to wait for the next tick.
See how Findable reports into the monitoring stack your team already runs.