Hi all! I have a questions regarding the max context iterations limit (https://github.com/infracost/infracost/blob/3d1d555e07082add91d6dd27425a9863a626038c/internal/hcl/evaluator.go#L64). Is there a particular reason for this value of 120? If I were to adjust this to say…360, would it create a rift in the universe?

Hi Terry, I don’t think there’s a specific reason for 120, but @mysterious-teacher-68276 may have a better idea when he’s back online. Look’s like it’s been increased before so if you want, submit a PR and we’ll see what he says.

thank you @crooked-daybreak-55253!

Yeah that’s not a bad idea. The way the parsing works is that it runs in a loop that expands (e.g.

for_each

) and evaluates blocks until it detects that nothing has changed. So the max is there to prevent it from looping infinitely.

ahh, gotcha

hi @wooden-notebook-80950 as Tim mentioned this variable is to prevent the contextual evaluation loop getting stuck and looping forever. However, in reality you should never really hit this variable. Are you actively seeing your projects hit this threshold? Or are you just inferring that from some other incorrect output. It’s most likely that if you are seeing this threshold hit, we’re hitting some evaluation bug in one or more of your projects which means we get stuck and have unchanged module contexts. So it would be good to share your specific case if you are seeing this and hopefully we can put in a fix. I would not recommend simply “bumping” the number of max context iterations as this will significantly impact performance evaluating projects. It’s much better to understand your specific case and fix the underlying bug there.

Thanks @mysterious-teacher-68276, yes my projects are actively hitting this threshold, at least I am assuming so since I’m seeing a number of errors related to hitting the max context iterations like so:

module.autoscale_sync_job_records.module.this_autoscale msg="hit max context iterations evaluating module module.autoscale_sync_job_records.module.this_autoscale"

This is followed but a somewhat strange error saying comments can no longer be posted to Github…

[2023-08-23T18:20:36.227Z] Error: template: markdown-html.tmpl:80:115: executing "markdown-html.tmpl" at <.PastBreakdown.TotalMonthlyCost>: nil pointer evaluating *output.Breakdown.TotalMonthlyCost
[2023-08-23T18:20:36.328Z] Error posting cost report to Github.

ok a couple of questions: 1. what version of infracost CLI are you using 2. what is this module:

module.autoscale_sync_job_records.module.this_autoscale

- does it use any 3rd party modules inside it or is it just vanilla terraform? 3. The comment posting might be related, but I don’t think max iterations should affect this

1. 0.10.28 2. That module is a submodule of of of the included modules in the project that handles ddb autoscaling policies. So in other words, we have the root module which calls another module which calls

resource_autoscale

in it’s module. 3. 👍

so

module.autoscale_sync_job_records.module.this_autoscale

doesn’t contain any module calls?

nope

the module layering should be fine, it’s likely that there is something within that module that’s causing an infinite contextual loop. Without being able to see the specifics I can’t do much to help. Is there any way you can share the code with us to help debug. We can start a private chat to do this?