Observability Overview

Upblit’s observability layer gives you distributed traces and structured logs for every HTTP request your instrumented applications handle — with zero configuration beyond adding the SDK middleware.

What Gets Collected

Traces

A trace represents a single request flowing through your system. It contains one or more spans:

• Root span — created by the middleware for every incoming HTTP request
• Service spans — created by sdk.service() for internal service calls
• External spans — created by sdk.call() for outbound HTTP calls

Each span records:

• traceId — shared across all spans in one request
• spanId — unique per span
• parentSpanId — links child spans to their parent
• requestMethod — e.g., controller:POST, service:getUserById
• requestURL — HTTP path (empty for internal spans)
• responseStatus — HTTP status code
• durationMs — wall-clock duration

Logs

A log entry is a structured message emitted by your application code. Log entries are correlated to traces via traceId.

Each log entry records:

• level — fatal, error, warn, info, debug
• message — the log text
• traceId — correlates to a trace (if emitted within a request context)
• timestamp — server-side receipt time
• clientTimestamp — SDK-side emission time

How Data Flows

Your App → SDK Middleware → In-Memory Buffer → POST /ingest/traces
                                             → POST /ingest/logs
                                                    ↓
                                             MongoDB (stored)
                                                    ↓
                                             Dashboard Query

Data is buffered in memory and flushed every 30 seconds (configurable). On flush failure, the batch is re-queued — no data is dropped.

Viewing Telemetry

Go to your project’s Observability page:

/dashboard/orgs/[orgId]/projects/[projectId]/observability

Telemetry and logs are fetched in parallel when the page loads.