When to Use causal-order

Good Fits

The library is a strong fit for:

• audit timeline reconstruction
• replay analysis
• multi-region debugging
• offline sync inspection
• late-arrival stream handling
• distributed incident analysis

It is especially useful when:

• the events come from multiple nodes or regions
• wall-clock timestamps are imperfect
• ordering claims need explanation
• concurrency matters
• weak metadata should not be silently normalized

Less Useful Fits

It may be unnecessary when:

• everything happens on one trusted process with one reliable sequence
• plain local ordering is already obvious
• the system does not need explainable event order

The Litmus Test

If your team has ever said:

• “we just sorted by timestamp”
• “this replay looks newer than the original”
• “the device synced later, so the timeline got weird”
• “we need to know whether this order is real or inferred”

then causal-order is likely solving a real problem for you.