draft operational-models poster #6
Conversation
|
|
||
| ## Evaluation approach | ||
|
|
||
| Once we've enumerated possible solutions then we will evaluate them to choose the best for us. The choice should balance: |
There was a problem hiding this comment.
Love how this poster explicitly lists the tradeoffs!
| - cost (lower cost solutions are better) | ||
| - understandability to lab members (more understandable solutions are better) | ||
|
|
||
| After that we will implement the chosen solution by creating a working infrastructure to "dry run" our flu_ar2 model and then judging how it went, both process and end product. Details: Write python code located in a https://github.com/reichlab/operational-models `pulumi` branch under a `infrastructure` dir (say) that runs this container: |
There was a problem hiding this comment.
I'm a bit confused about the Pulumi reference here, since infrastructure as code is only one of the possible solutions listed above. Is this saying that we'll try out Pulumi as a way to get more information to make the decision?
| - `CHANNEL_ID=C02UELXDPQQ` # "#tmp" channel | ||
| - `DRY_RUN=1` # enabled | ||
|
|
||
| ## Current infrastructure |
There was a problem hiding this comment.
This is great information that can help inform our assessment of your proposed solutions. Am suggesting alternate heading names to align with the "validation" part of the project posted template (which also opens the door for a "what do we need to answer" section.
| ## Current infrastructure | |
| ## Validation | |
| ### What do we already know? | |
| #### Current infrastructure |
| @@ -0,0 +1,60 @@ | |||
| # Project Poster: Use infrastructure as code (IaC) for the lab's operational-models | |||
There was a problem hiding this comment.
As has happened on 100% of these so far, I have a question about scope where the answer may be, "the scope listed in this title is appropriate". But when I was thinking about this, I was thinking of something like "Use IaC to manage the lab's AWS resources", where those resources include stuff related to the lab's operational-models, but also a couple of one-off S3 buckets that are used to store data:
- infectious-disease-data
- covid-clade-counts
I similarly manually created those buckets (with Becky's help) and set up various configurations for them by navigating the AWS console, editing JSON configs, etc. It'd be great to eventually have a better process for creating buckets like this as well.
I've done my best to incorporate discussion on the slack thread. I took the template as inspiration for my headings - the ones here are not perfect, and I'm open to their being changed. I just needed to get something down. There are a few TODO items. Thanks!