PR is too large for us to process being shown for most MRs

[echarrod]

Describe the bug

AI code review reports:

Uh oh! That's a big one.
This PR is too large for us to process, we gather the full context, including all file contents before and after the changes (not just the diffs), plus metadata. In general, keeping PRs small is a best practice.
ERROR: Request body size is 1.62 MB, which exceeds the 1MB limit.
Added by gitStream

This is now happening for most MRs since we updated index.docker.io/gitstream/rules-engine from 1.0.121 -> 1.0.130 yesterday

Config:

manifest:
  version: 1.0

config:
  ignore_files:
    - 'go.sum'
    - 'go.work.sum'
    - '*.pb.go'
    - '*.pb.gw.go'
    - 'bin/**'
    - 'static/**'
    - 'blobs/**'

automations:
  # AI code review for new changes
  linearb_ai_review:
    if:
      - {{ not pr.draft }}
      - {{ not is.bot }}
    run:
      - action: code-review@v1
        args:
          approve_on_LGTM: {{ approve_pr_on_lgtm }} # optional arg, you can remove it
          guidelines: {{ guidelines }}

# Define variables
# Add conditions for PR approvals. For example - allow approval only for specific users
approve_pr_on_lgtm: false

# Custom expressions for bot detection
is:
  bot: >-
    {{ branch.author | match(list=[
        '_bot', 
        '[bot]'
    ]) | some }}

# Add `guidelines` to add your prompts to the review, you can have org wide instructions
# by placing the `REVIEW_RULES.md` in the `cm` repo or team level instructions by adding it to
# the team repo.
# Note: To when the file is in the `cm` repo, use this example: `../cm/REVIEW_RULES.md`
guidelines: {{ "./.cm/REVIEW_RULES.md" | readFile() | dump }}

Expected behavior

• This was happening occasionally before, now it's for most MRs

Screenshots

Additional context

Running on self-hosted GitLab instance, using image index.docker.io/gitstream/rules-engine:1.0.130

[vim-zz]

We are increasing the limit today to 5MB,
if in few hours you still experience this, let me know

[echarrod]

We are increasing the limit today to 5MB, if in few hours you still experience this, let me know

Thanks. Anecdotally from reviewing MRs though it seems like the size calculation has changed in between 1.0.121 and 1.0.130.

When we upgraded to 1.0.130, we got this a lot more frequently.
And then when we downgraded to 1.0.121 we get it a lot less again.

Even for the same MR without any changes, 1.0.130 said it was too large, but then 1.0.121 (after we rolled back) parsed it fine.

I appreciate increasing the limit, but just also wanted to flag that I think there's been a regression in how this limit is calculated, or at least a lot of extra information is being included for the same MR