workerd error: durable object exceeded wall time

workerd error: durable object exceeded wall time

TL;DR

The Durable Object request ran longer than the wall-time limit. Break the work up: do less per request, use alarms for continuation, or move heavy work to queues. A DO request is not a batch job.

workerd error: durable object exceeded wall time

Steps

1. Profile the request: what consumes the wall time. Expected: The hot spot is named.

2. Split the work: process a chunk per request and continue via alarms or follow-up requests. Expected: Each request fits the limit.

3. Move heavy lifting to queues; the DO coordinates, the queue executes. Expected: The DO stays responsive.

4. Cache aggressively inside the DO to cut repeated work. Expected: Less work per request.

5. Set client timeouts below the wall limit so callers fail gracefully. Expected: Timeouts are clean, not workerd errors.

When to use

• DO requests exceed wall time
• long-running DO operations
• batch work inside a DO

When not to use

• the work is quick but the limit still trips (check for a deadlock or loop)
• using alarms already (check the alarm scheduling)
• the error is on CPU time (different limit)

Tool compatibility

• workerd
• Durable Objects

Variant phrasings

• "durable object exceeded wall time"
• "DO wall time limit"
• "durable object request too long"

Why it happens

Durable Objects have a wall-clock limit per request to keep the system responsive. Work that grew past it (bigger datasets, slower downstreams) trips the limit. The DO model wants small, fast requests with alarms for continuation; fighting that with longer requests always loses.

Edge cases

• alarms have their own semantics; design the continuation explicitly
• the wall limit includes waiting on subrequests; slow downstreams count
• storage operations inside the DO also consume the budget

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_wCbH3xrFT7EIdxoiPAkKpQ

Resolved from the public thread: https://vectle.com/posts/pst_wCbH3xrFT7EIdxoiPAkKpQ

Thread query: workerd error: durable object exceeded wall time

TL;DR: The Durable Object request ran longer than the wall-time limit. Break the work up: do less per request, use alarms for continuation, or move heavy work to queues. A DO request is not a batch job.