The shortest path
to useful context.
Capture less. Understand more. Trace the journey from a request event to a deliberate, well-defined handoff.
Explore the event path ↗Every connection
needs a contract.
What enters.
Choose route templates and stable event fields. Give the collector only the context it needs.
What stays.
Bound the buffer by capacity and age. A slow destination should not quietly consume the application.
What travels.
Keep an event identifier through retries. Explain duplication and delivery boundaries before depending on them.
Notes from the boundary.
Design a compact event envelope
The fields that make distributed request events useful.
Read the guide ↗02 / APIPulse EdgeBuffering and backpressure
A collector should have predictable behaviour under pressure.
Read the guide ↗03 / APIPulse EdgeDelivery, duplicates and restarts
What a receiving service can safely assume.
Read the guide ↗A few useful answers
Is this the APIPulse API health workflow?
This resource focuses on edge collection and transport. API-level interpretation requires the receiving system to apply its own operation contracts.
Should every request be collected?
Only when the operational need and data budget justify it. Sampling decisions should preserve important errors and remain visible in downstream analysis.
Can events contain request bodies?
Prefer a small allowlist of metadata. Request bodies can contain secrets or personal data and require a separate, justified handling policy.
What happens during a disconnect?
That depends on the chosen buffer and overflow policy. The guides explain how to make that behaviour bounded and observable.