floci-io/floci · error · AwsException

ResourceInUse

ResourceInUse

Error message

Auto Scaling group '{name}' has {count} instance(s). Set ForceDelete=true to delete anyway.

What it means

Thrown by deleteAutoScalingGroup when the group still has instances not in the Terminated lifecycle state and ForceDelete is not set. Amazon EC2 Auto Scaling refuses to delete non-empty groups to prevent orphaned instances; ForceDelete=true terminates the members first. Floci reproduces this as ResourceInUse (HTTP 400), matching AWS.

Source

Thrown at src/main/java/io/github/hectorvent/floci/services/autoscaling/AutoScalingService.java:282

        if (minSize != null) { asg.setMinSize(minSize); }
        if (maxSize != null) { asg.setMaxSize(maxSize); }
        if (desiredCapacity != null) { asg.setDesiredCapacity(desiredCapacity); }
        if (defaultCooldown != null) { asg.setDefaultCooldown(defaultCooldown); }
        if (availabilityZones != null) { asg.setAvailabilityZones(new ArrayList<>(availabilityZones)); }
        if (subnetIds != null) { asg.setSubnetIds(new ArrayList<>(subnetIds)); }
        if (healthCheckType != null) { asg.setHealthCheckType(healthCheckType); }
        if (healthCheckGracePeriod != null) { asg.setHealthCheckGracePeriod(healthCheckGracePeriod); }
        if (terminationPolicies != null) { asg.setTerminationPolicies(new ArrayList<>(terminationPolicies)); }
        groups.put(asgKey(region, name), asg);
    }

    public void deleteAutoScalingGroup(String region, String name, boolean forceDelete) {
        AutoScalingGroup asg = requireGroup(region, name);
        List<AsgInstance> active = asg.getInstances().stream()
                .filter(i -> !"Terminated".equals(i.getLifecycleState()))
                .collect(Collectors.toList());
        if (!active.isEmpty() && !forceDelete) {
            throw new AwsException("ResourceInUse",
                    "Auto Scaling group '" + name + "' has " + active.size()
                            + " instance(s). Set ForceDelete=true to delete anyway.", 400);
        }
        if (forceDelete && ec2Service != null && !active.isEmpty()) {
            active.stream()
                    .map(AsgInstance::getInstanceId)
                    .filter(Objects::nonNull)
                    .forEach(instanceId -> {
                        try {
                            ec2Service.terminateInstances(region, List.of(instanceId));
                        }
                        catch (AwsException ignored) {
                            // ForceDelete should remove stale ASG membership even if EC2 no longer has the instance.
                        }
                    });
        }
        groups.remove(asgKey(region, name));
        // clean up associated hooks and policies

View on GitHub (pinned to 62ff490619)

Solutions

Set desired capacity to 0 (UpdateAutoScalingGroup with DesiredCapacity=0), wait for instances to reach Terminated, then delete
Or set ForceDelete=true on the DeleteAutoScalingGroup call, which terminates instances and deletes the group
For graceful teardown: detach instances with ShouldDecrementDesiredCapacity=true, wait, then delete

Example fix

// before
DeleteAutoScalingGroupRequest req = DeleteAutoScalingGroupRequest.builder()
    .autoScalingGroupName(asgName)
    .build();

// after
DeleteAutoScalingGroupRequest req = DeleteAutoScalingGroupRequest.builder()
    .autoScalingGroupName(asgName)
    .forceDelete(true)
    .build();

Defensive patterns

Strategy: validation

Validate before calling

Group g = autoscaling.describeAutoScalingGroups(DescribeAutoScalingGroupsRequest.builder()
        .autoScalingGroupNames(asgName).build())
    .autoScalingGroups().get(0);
long active = g.instances().stream()
    .filter(i -> !"Terminated".equals(i.lifecycleState())).count();
if (active > 0) {
    // scale in first, or set forceDelete(true) on the delete request
}

Try / catch

catch (ResourceInUseException e) { log.warn("ASG {} not empty", asgName); retryWithForceDelete or scaleInFirst(); }

Prevention

Scale groups to DesiredCapacity=0 before deletion in teardown scripts
Treat ResourceInUse as a retryable-with-action signal, not a hard failure

When it happens

Trigger: Calling DeleteAutoScalingGroup with the group's AutoScalingGroupName while desiredCapacity > 0 or any instance is InService/Pending/Standby, without setting ForceDelete=true.

Common situations: Tearing down test stacks without scaling the group to zero first; cleanup scripts that assume delete is unconditional; instances mid-launch from a recent scale-out; standby instances that never terminated.

Related errors

AI-assisted analysis of floci-io/floci@62ff490619 (2026-08-14). Data as JSON: /api/errors/869e2876f5c98ace. Report an issue: GitHub.