Logging & observability

Все три сервиса (Web, MCP, CtraderCliNode) логируют через Serilog как compact JSON на stdout — container runtimes и log collectors (Loki, ELK, CloudWatch, Azure Log Analytics, Datadog) absorb структурированные события directly, нет free-text parsing.

Pipeline

• Sink 1 — Console (compact JSON). RenderedCompactJsonFormatter; каждое событие носит full OpenTelemetry resource identity — service.name (cmind-web / cmind-mcp / cmind-node-agent / cmind-copy-agent), service.version, service.namespace (cmind), deployment.environment — плюс trace_id / span_id из ambient Activity (ActivityEnricher) и LogContext scopes. Trace ids позволяют CloudWatch Logs Insights и Azure Log Analytics pivot log↔trace даже с нет collector deployed.
• Sink 2 — OTLP (optional). Когда OTEL_EXPORTER_OTLP_ENDPOINT set, логи также export через OTLP с то же resource attributes, наряду с OpenTelemetry metrics и traces (ASP.NET Core, HttpClient, runtime instrumentation) из AddAppTelemetry.
• Sink 3 — Azure Monitor (optional). Когда APPLICATIONINSIGHTS_CONNECTION_STRING set, traces и metrics export natively to Application Insights (AddAzureMonitorTraceExporter / AddAzureMonitorMetricExporter) — нет collector. See cloud-native секция ниже.
• Source-generated messages. App logs используют strongly-typed LogMessages extensions (Core/Logging/LogMessages.cs) с stable EventIds — никогда ad-hoc ILogger.LogInformation.
• Request logging. UseSerilogRequestLogging() emits один structured summary per HTTP request (method, path, status, elapsed ms).

Configuration

Levels tunable per service через Serilog секцию в appsettings.json (read через ReadFrom.Configuration), например:

{
  "Serilog": {
    "MinimumLevel": {
      "Default": "Information",
      "Override": { "Microsoft.AspNetCore": "Warning", "Microsoft.EntityFrameworkCore": "Warning" }
    }
  }
}

Override во время выполнения с env vars, например Serilog__MinimumLevel__Default=Debug.

Shipping на collector

Set один env var per service, point на OTLP endpoint:

OTEL_EXPORTER_OTLP_ENDPOINT=http://otel-collector:4317

• Helm: --set observability.otlpEndpoint=http://otel-collector:4317.
• Compose / cloud: add env var к каждому service.

Из collector, fan out to backend (Tempo/Jaeger traces, Prometheus metrics, Loki logs) с trace↔log correlation intact.

Cloud-native backends (нет extra collector)

Оба managed deployments wired для native observability stack out of box — нет OTLP collector needed.

Azure — Application Insights + Log Analytics

deploy/azure/main.bicep provisions workspace-based Application Insights component, passes его connection string к Web и MCP как APPLICATIONINSIGHTS_CONNECTION_STRING. Result:

• Traces + metrics flow natively в App Insights (Application Map, live metrics, end-to-end transaction search), correlated by trace_id.
• Logs (compact JSON на stdout) land в то же Log Analytics workspace через Container Apps appLogsConfiguration, поэтому AppTraces / ContainerAppConsoleLogs_CL join на trace id.
• Set optional otlpEndpoint Bicep param для также fan out на collector.

Query логи в Log Analytics (JSON line в Log_s):

ContainerAppConsoleLogs_CL
| extend log = parse_json(Log_s)
| where log.["service.name"] == "cmind-web"
| project TimeGenerated, level = log["@l"], msg = log["@m"], trace_id = log.trace_id

AWS — X-Ray + CloudWatch (ADOT sidecar)

deploy/aws/main.tf runs AWS Distro для OpenTelemetry (ADOT) collector как sidecar в каждой Fargate задаче. App exports OTLP на http://localhost:4317; sidecar ships:

• traces → AWS X-Ray (awsxray exporter),
• metrics → CloudWatch (awsemf, namespace cmind, log group /ecs/<prefix>/metrics).
• Logs stay на awslogs driver как compact JSON; CloudWatch Logs Insights auto-discovers JSON fields, поэтому вы filter / stats на trace_id, service.name, @l, и т.д.

Task role (aws_iam_role.task) carries AWSXRayDaemonWriteAccess + CloudWatchAgentServerPolicy.

Query логи в CloudWatch Logs Insights:

fields @timestamp, @l, @m, trace_id, `service.name`
| filter `service.name` = "cmind-web"
| sort @timestamp desc

Health endpoints (также используется probes)

Endpoint	Service	Meaning
/alive	Web	Liveness — process only.
/health	Web	Readiness — includes database check.
/version	Web, MCP	Product + protocol version (MCP liveness/readiness).

Mapped во все environments (ранее Dev-only) поэтому Kubernetes и cloud probes работают в production.