The Three Pillars of Observability, Wired Together with OpenTelemetry

Monitoring vs. Observability

Monitoring answers questions you predicted (“is CPU > 80%?”). Observability lets you ask questions you didn’t predict (“why is the p99 for checkout slow, but only for EU users on mobile?”). Getting there means correlating the three signals instead of treating them as separate tools.

The Three Pillars

• Metrics — cheap, aggregatable numbers over time. Great for dashboards, SLOs and alerting (the RED method: Rate, Errors, Duration).
• Logs — discrete, high-detail events. Great for the what happened once you know where to look.
• Traces — the request’s journey across services, with timing per hop. This is the connective tissue that tells you where to look.

OpenTelemetry Ties Them Together

The superpower is shared trace context. When a log line, a metric exemplar and a span all carry the same trace_id, you can pivot from a slow trace straight to the exact logs for that request.

from opentelemetry import trace
tracer = trace.get_tracer(__name__)

with tracer.start_as_current_span("charge_card") as span:
    span.set_attribute("payment.provider", "stripe")
    result = provider.charge(amount)
    span.set_attribute("payment.status", result.status)

Practical Advice

1. Start with auto-instrumentation, then add manual spans only around the business logic you actually care about.
2. Run the OpenTelemetry Collector so you can change backends without touching app code.
3. Use tail-based sampling in production — keep the slow and errored traces, drop the boring ones.