Workers Paid Workflows now default to seven‑day instance retention

Effective September 10, 2026, new Workflows created on the Workers Paid plan keep completed and errored instance data for only seven days by default, down from the previous thirty‑day default. The change is intended to lower storage costs, while the maximum retention window remains at thirty days.

Why the retention window matters

Engineers who rely on Workflows for long‑running data pipelines, audit trails, or debugging need to know how long instance state is available. A shorter default means that, unless you explicitly extend it, logs and result payloads will disappear after a week, which can affect post‑mortem analysis, compliance reporting, and any downstream processes that pull historical data.

Configuring retention per workflow or instance

You can override the seven‑day default by supplying successRetention and/or errorRetention when you create a workflow instance. The same fields are available in the dashboard for per‑workflow settings. The API call looks like this:

const instance = await env.MY_WORKFLOW.create({
  retention: {
    successRetention: "2 days",
    errorRetention: "30 days",
  },
});

Both values accept human‑readable durations (e.g., "2 days", "30 days"). The platform enforces the overall maximum of thirty days, so any value beyond that is capped.

Operational considerations

With a seven‑day default you should review any monitoring or alerting that depends on instance state older than a week. If you have automated clean‑up jobs, they may no longer need to run as aggressively. Conversely, if you need longer visibility, incorporate explicit retention settings into your deployment scripts or CI pipelines to avoid accidental data loss.

Because the Workers Free plan still defaults to three days, the gap between free and paid tiers widens. Teams that run mixed workloads should verify that their free‑tier jobs are not unintentionally missing critical history.

Security and compliance angle

Shorter default retention reduces the amount of data stored at rest, which can lower the attack surface and simplify compliance with data‑retention policies. However, if your organization mandates longer audit logs, you must explicitly set the retention values to meet those requirements; otherwise the default will cause premature deletion.

Related CloudNinjas coverage: hands-on guides.

What This Means For Practitioners

Audit your existing workflow creations and add explicit successRetention or errorRetention parameters where you need more than seven days of history. Update any monitoring, alerting, or compliance checks that assume a thirty‑day window. Finally, verify that the dashboard settings for each workflow align with your retention policy to avoid surprises.