Home › Key concepts › State and durable waits
Key concept · Run (advanced) · the same on every platform
State and durable waits
Remember where a long task was, survive crashes, and wait days for a human.
What it is
Real tasks pause: for a customer's reply, a manager's approval, a document. Durable execution saves each step so that after a crash or a three-day wait the task continues from where it stopped, without redoing finished steps.
Why it matters
Without it, a restart can repeat a payment or lose a half-finished claim.
An example
A claim waits two days for a photo; when it arrives, the workflow resumes at step 4, not step 1.
The common mistake
Keeping agent state only in memory.
Where you meet it on each platform
The idea stays the same; only the product name changes.
| Job | AWS | Azure | Google Cloud | Databricks | Open source |
|---|---|---|---|---|---|
| Agent runtimewhere a long-running agent keeps its state and waits safely | Bedrock AgentCore | Foundry Agent Service / Durable Functions | Vertex AI Agent Engine | Mosaic AI Agent Framework | DBOS or Temporal + your loop |
Related ideas
Where it shows up
Part of the Run level on these paths: AI Engineering (tools and agents).