Start at the workload boundary
Prefer stable service, workload, queue, and route names over ephemeral pod IDs, raw resource paths, or provider payloads.
Selection framework
Choose the boundary that knows the result
Pair resource pressure with demand
CPU, memory, restarts, lag, and throttling need request or job volume plus release context before they support a capacity decision.
Retain the specialist signal
Use infrastructure, tracing, and provider systems for their native detail. Send selected application outcomes to Telemetry for SQL analysis and customer impact.
Shared event contract
Keep the first schema interoperable
- service, region, environment, release, status, and duration_ms
- provider, runtime, workload, queue, dependency, and controlled error_type
- request volume, resource saturation, retry count, and customer-impact fields
Adapt these fields to the workflow, but keep units, status categories, identifiers, and ownership explicit before several services depend on the same table.
Verification
Exercise failure paths before alerts
- 1Exercise one healthy deployment and one controlled provider or workload failure.
- 2Confirm region, release, service, and runtime dimensions remain bounded.
- 3Test retry and timeout behavior without recursive telemetry failures.
- 4Compare infrastructure changes with request, job, or product outcomes.
Implementation guides
Match the framework, provider, or runtime
Sourced implementation guide
Cloudflare Worker Telemetry
Log edge request status, latency, provider failures, cron runs, and queue workers from Cloudflare Workers.
Open guideSourced implementation guide
AWS Lambda Structured Event Monitoring
Track Lambda invocations, cold starts, duration, retries, and business outcomes with compact structured events.
Open guideSourced implementation guide
AWS ECS and Fargate Telemetry Integration
Add ECS task family, revision, availability zone, and release context to application outcomes using the local task metadata endpoint.
Open guideSourced implementation guide
AWS SQS Queue Observability
Track SQS message age, receive attempts, processing outcomes, batch behavior, and dead-letter routing without storing message bodies.
Open guideSourced implementation guide
Google Cloud Run Telemetry
Track Cloud Run request and job outcomes, cold-start context, instance concurrency, retries, latency, and releases with application-owned structured events.
Open guideSourced implementation guide
Google Cloud Pub/Sub Telemetry
Track Pub/Sub publish, delivery, acknowledgement, redelivery, ordering, queue age, and dead-letter outcomes with stable message identifiers.
Open guideSourced implementation guide
Azure Functions Telemetry
Track Azure Functions HTTP, timer, queue, and event-trigger outcomes with invocation, retry, latency, release, and customer-impact context.
Open guideSourced implementation guide
Azure Service Bus Telemetry
Track Azure Service Bus sends, receives, lock renewal, settlement, redelivery, deferral, and dead-letter outcomes without collecting message bodies.
Open guideSourced implementation guide
Kafka Consumer Lag and Processing Analytics
Monitor Kafka consumer outcomes, partition lag, processing latency, retries, and poison-message handling with structured events.
Open guideSourced implementation guide
Kubernetes Workload Telemetry
Send bounded Kubernetes restart, readiness, and rollout observations to Telemetry for workload-level SQL analysis.
Open guideSourced implementation guide
Amazon Bedrock Model Telemetry
Measure Bedrock Converse requests by model or inference profile, token usage, latency, stop reason, retries, and business outcome.
Open guideSourced implementation guide
Azure OpenAI Responses Telemetry
Track Azure OpenAI deployment usage, token cost, latency, failures, and product outcomes with editable SQL-ready events.
Open guideSourced implementation guide
Google Gemini API Telemetry
Track Gemini request latency, model usage, tokens, finish reason, errors, and reviewed product outcomes without storing prompts or generated content.
Open guideSourced implementation guide
OpenTelemetry Context in Structured Events
Connect Telemetry business events to existing OpenTelemetry traces by recording safe trace and span identifiers at workflow boundaries.
Open guideVerify one end-to-end workflow first
Create a free API key, choose the closest runtime guide, and send one known success and failure before expanding coverage.