EcsContainerStopTimeout
Custom timeout duration for stopping ECS containers on Rack instances. This parameter defines the time (in seconds) ECS waits after sending a SIGTERM before issuing a SIGKILL, allowing for graceful shutdowns.
By default, this value is unset, meaning ECS will use its default 30-second stop timeout or any custom configuration already set at the ECS level.
| Default value | "" |
| Allowed values | Numerical values in seconds (e.g., 10, 60, 120) |
Use Cases
- Increase the timeout for applications that need additional time to drain active connections, flush caches, or complete in-flight requests during shutdown.
- Set a shorter timeout for stateless workers that can be stopped quickly, reducing deployment time.
- Tune shutdown behavior for applications with complex cleanup processes, such as closing database connections or persisting session state.
Additional Information
When ECS stops a container, it first sends a SIGTERM signal to the container's main process. The container then has the configured timeout period to shut down gracefully before ECS sends a SIGKILL to force termination. If your application does not handle SIGTERM, it will be forcibly killed after the timeout expires.
This parameter maps to the ECS_CONTAINER_STOP_TIMEOUT ECS agent configuration variable and is written to /etc/ecs/ecs.config on each EC2 instance. Because it is an agent-level setting, it applies to all containers on the instance, not individual services. This parameter has no effect on Fargate tasks, which use the stopTimeout setting in the task definition instead.
For more details, see the ECS Agent Configuration documentation.
$ convox rack params set EcsContainerStopTimeout=60
See Also
- EcsPollInterval
- Rack Parameters for a full list of available parameters
