Hydra tightens L2 UTxO handling to protect large-head close timing

What happened

Hydra merged PR #2837, which fixes #2831 by making Layer 2 UTxO values strict when they are constructed. The change uses forced TxOut values and UTxO handling so deferred computations, known as thunks, do not accumulate until a close transaction needs them.

The source record says that, on large heads, forcing those accumulated values during close construction could add more than a second of delay. That could push the close transaction beyond its on-chain validity window.

Why it matters

A Hydra head close is the step that settles Layer 2 funds back to Layer 1. Reducing a construction stall at that point is therefore an operator-facing liveness improvement for large heads, where timing against the close window matters.

The wider strictness work also covers UTxO paths such as JSON and CBOR decoding and deposit observation. The record says it is backed by nothunks property tests and benchmarks, though an earlier broad approach showed performance regressions before a targeted adjustment reversed them.

What to watch

The next useful receipt is a tagged Hydra release containing #2837, followed by evidence from large-head timing or benchmark results. Teams extending hydra-node can also watch whether the new strictness patterns and tests are carried into their own UTxO-handling paths.