AWS Spot instances look ideal for running our workers in the lower environments. Does Octopus have any support for Spot instances? Or any documents that would help set it up ?
We don’t currently have any built-in support for integrating workers with AWS spot instances.
Wiring this up yourself may be possible, depending on exactly what you are looking to achieve. Though this isn’t something we’ve done ourselves (yet).
Perhaps using an auto-scale group, and the register-worker and deregister-worker tentacle.exe commands.
One catch would be that you would have to maintain at least one worker in the pool (at least at times that deployments may happen) because we don’t currently expose a hook for you to know when a deployment needs a worker. The deployment will fail if no workers are present.
We’d be interested to hear any thoughts you have on what your ideal solution would look like, and what you would need to implement this?
First-class support for spot instances isn’t on our short-term radar.
We are currently working on implementing hosted worker pools. These will be dynamic workers hosted by us. These will initially be offered to Octopus Cloud customers, but will likely later be offered to self-hosted customers also.
Thanks for the reply, I will look into it and will update this post if I get much traction.