Hello! I was wondering if there was an easy way to wait for a newly created

aws.ec2.Instance

to be in a running state before attempting to do a

command.remote.Command

on that host.

The normal way to do this is to write a custom "wait script" as part of an apply. Found an k8s example of this at https://gist.github.com/lukehoban/fd0355ed5b82386bd89c0ffe2a3c916a But the idea's similar for any resource pair. Take an output from the ec2.Instance and "apply" it to run some logic in a while loop that doesn't return till it's ready. Then exit the while loop and return a value to resolve the output. If the command resource then depends on that output it won't run till the ec2 instance is ready.

hmm, so do you think something like this would work:

async function waitForInstance(id: string) {
  if (pulumi.runtime.isDryRun()) {
    return;
  }

  // Wait for up to 10 minutes
  for (let i = 0; i < 60; i++) {
    let instance = await aws.ec2.getInstance({ instanceId: id });
    if (instance.instanceState === "running") {
      return instance;
    }
    pulumi.log.info(`Waiting for Instance to start (${id})`);
    // Wait for 10s between polls
    await new Promise((r) => setTimeout(r, 10000));
  }

  throw new Error("timed out waiting for Instance to start");
}

Yes I think so, I don't think repeated getInstances for the same ID are cached so I suspect this would work