What happens when an actor fails?

When an Apify actor fails, the run terminates with a FAILED status and the platform records the error details, exit code, and run log. Understanding why actors fail — and how to diagnose failures quickly — is critical for maintaining high success rates and avoiding maintenance flags that reduce your Store visibility. The most common failure categories are: runtime errors (unhandled exceptions in your code), timeouts (actor exceeds its configured timeout or the default 3600 seconds), out-of-memory crashes (actor tries to use more RAM than allocated), network errors (target website is down or blocks requests), and input validation failures (user provides invalid or missing required inputs). Each category produces different error signatures in the run log. When a failure occurs, here is what happens on the platform side. The run status is set to FAILED. The error message and stack trace are captured in the run log. If the actor was mid-way through producing results, any data already pushed to the dataset is preserved — you do not lose partial results. The user who triggered the run is charged only for the compute time actually consumed, not the full expected duration. If the failure happened before any meaningful work, the charge is minimal. For actor developers, failures directly impact your quality score and Store visibility. Apify tracks your actor's success rate over the past 7 days. If the rate drops below a threshold (typically around 80%), your actor may receive a maintenance flag. This is why catching failures before they reach production — using the ApifyForge Test Runner, Schema Validator, and Cloud Staging — is so valuable. A single automated test that costs $0.001 can prevent a failure that costs you weeks of reduced Store visibility. ApifyForge monitors failure rates across your entire fleet through the dashboard. You can see which actors are failing most frequently, identify patterns in failure types, and prioritize fixes based on revenue impact. For step-by-step debugging instructions, see the related question about how to debug a failed actor run. For testing tools that prevent failures, see the questions about the Test Runner and Cloud Staging.