--- cuppy_hub: azure cuppy_path: changelog/2026-10-21-observability-family.md cuppy_section: docs cuppy_doc: KHY7-Z9PW cuppy_doc_uuid: 2da6f106-942c-4afa-9c8c-479a71d9d7b7 cuppy_permalink: https://cuppy.io/azure/d/KHY7-Z9PW cuppy_version: 1 cuppy_updated: 2026-10-23T16:16:52.575Z cuppy_canonical: https://cuppy.io/azure/docs/changelog/2026-10-21-observability-family cuppy_mcp: https://cuppy.io/mcp cuppy_body_sha256: 2d5d5a8ab69518554e5efb3817d8d05352951651f6a4a7ceb03dd6eaf8f4e66d # cuppy_body_sha256 is over the document body ALONE, excluding both generated blocks. --- # Azure hub — family 7, Observability (2026-10-21) ## What's in it 7 new documents including this note, bringing the hub to 108: - `reference/monitor/what-to-use.md`: which signal goes where, and what's retired (classic Application Insights, the Log Analytics agent, the Azure Diagnostics extension, the HTTP Data Collector API, diagnostic settings storage retention, the `@azure/monitor-query` package). - `guides/monitor/instrument-node-app-with-opentelemetry.md`: workspace and Entra-only Application Insights in Bicep, the Microsoft OpenTelemetry Distro with a managed identity, sampling, custom spans and metrics. - `reference/monitor/application-insights-errors.md`: ingestion responses (400, 401, 403), network causes, sampling, daily cap and size limits, duplicate distros. - `reference/monitor/log-analytics-workspaces.md`: workspace design, table plans, retention, access modes and roles, daily cap, queries with `@azure/monitor-query-logs` and the query limits. - `reference/monitor/diagnostic-settings-and-data-collection.md`: diagnostic settings rules and errors, DCR limits, Azure Monitor Agent identity, the Logs Ingestion API with a custom table, a `Direct` DCR and TypeScript. - `reference/monitor/alerts.md`: rule types, limits and rate limits, a metric alert and a log search alert with its own identity, why rules don't fire or fail to save, action group HTTP errors. ## 2026 changes captured - **`@microsoft/opentelemetry`** (1.5.0, Node.js 22+) is the new Microsoft OpenTelemetry Distro; by default it keeps **5 traces per second**, so set `tracesPerSecond: 0` to use a ratio. `@azure/monitor-opentelemetry` still works; never load both. - **`@azure/monitor-query` is deprecated**: use `@azure/monitor-query-logs` and `@azure/monitor-query-metrics`. - The **HTTP Data Collector API** needs TLS 1.2 from March 1, 2026 and reaches end of support on September 14, 2026; the Azure Diagnostics extension retired on March 31, 2026. - DCRs created since March 31, 2024 carry their own `logsIngestion` endpoint: a data collection endpoint is only needed for private link or older DCRs. ## How it was checked `MicrosoftDocs/azure-monitor-docs@14cccbe2` (CC BY 4.0 docs, MIT code) and the Azure Monitor troubleshooting articles in `SupportArticles-docs@ff674f85`; REST specs at `azure-rest-api-specs@c783b658`; Bicep builds and lints clean with 0.48.1 (workspace 2026-03-01, metric alerts 2026-01-01, scheduled query rules 2026-03-01, DCRs 2024-03-11, action groups 2023-01-01, diagnostic settings 2021-05-01-preview, its only version); TypeScript compiles under `tsc --strict` with `@microsoft/opentelemetry` 1.5.0, `@azure/monitor-query-logs` 1.0.0 and `@azure/monitor-ingestion` 1.2.0; CLI commands checked against 2.91.0 help (DCR commands are in the `monitor-control-service` extension). ## Still missing Workbooks and dashboards, Azure Monitor managed service for Prometheus and Managed Grafana (AKS child hub), metrics queries from TypeScript, Network Watcher flow logs, private link scopes (AMPLS) design, Container insights, cost optimisation for ingestion, KQL as a language. --- You are reading one file. The same content is free over MCP at https://cuppy.io/mcp with no signup: ranked search across all 107 documents in this hub, the merged view when you work in a layer, known workarounds matched to the error you hit, and depth control so you spend tokens on what you need. Start with the `_hello` tool. How to connect: https://cuppy.io/azure/ Something wrong in this document? Report it with the `session feedback` tool on the same endpoint.