use 'probe' outputs in 'actions'

Hi guys,

As far as I can understand the general workflow of the tool is:

• check steady state hypothesis
• run actions
• check steady state hypothesis again

In this flow probes represent steady state hypothesis and are not used for anything else.

For the use cases I see it would be extremely useful to use outputs of ‘probes’ as parameters to actions and have some conditional logic in between.

Example use case: Having hundreds of instances in AWS Region/AZ I don’t want to mess up with all of them as some instances/ASGs are not built for HA. So, I want to filter ASGs that:

• match certain tags (easy)
• have more than X healthy instances registered to load balancer and serving traffic (not so easy)
• pass resulting list of ASGs to action so that it will operate with this subset of instances choosing random ones to stop

I cam imagine doing all the checks necessary in action itself but this will lead to code duplication as I can see some of checks I’d like to do in probes already and naturally want to reuse existing code.

Regarding conditional logic (like, don’t stop instances in ASG if it has less than 2 healthy instances), do you think it is possible to integrate this to experiment yaml/json or the best place for this is action itself and one could pass conditions as parameters to action?

Please let me know your thoughts on this subject.

[Lawouach]

Hey folks,

Thank you @0rc for the awesome proposal. I think it certainly echoes what others have suggested in the past.

Russ and I knew initially we would need to explore one model along with the community so that the “right” approach would gradually emerge. Using a very declarative, sequential, model was a great starting point for its simplicity and ease to be understood and tried in various ways. I think it’s fair to say, Russ and I aren’t surprised it could evolve 😃

I see at least two features here:

• Being able to propagate generated/pulled data
• Offering more control inside the declaration

While the former is needed for the latter, I believe they can be both considered (and added) for their own merits.

My gut feeling, for now, is that the former is a great step forward. I’m almost tempted to say it could be included in 1.0 should we find the right format.

On the other hand, providing more control to the declaration requires careful thinking. As soon as we’ll add something, we may have to go all the way of a full blown language and I’m a little afraid of where that could lead us.

So, I am not advert of adding pattern matching and stop conditions but it would be great if we could keep some declarative level (rather than imperative style). For instance, if we could declare “guards” much like we declare “pauses”, that would have my favour.

At least for now.

[github-actions[bot]]

This Issue was closed because it was not reactivated after 7 days of being marked Stale.