Skip to content

ci: Issue Triage workflows fail with Argument list too long #26602

@cocosheng-g

Description

@cocosheng-g

What happened?

Both the Gemini Scheduled Issue Triage and Gemini Automated Issue Triage workflows are failing with an Argument list too long error. This occurs because the workflows load potentially huge JSON arrays or text of untriaged issues (which include issue bodies) directly into bash environment variables during their respective issue fetching and preparation steps.

What did you expect to happen?

The workflows should be able to safely fetch and process untriaged issues without exceeding the operating system's maximum environment variable size limits, enabling the Gemini issue analysis steps to process them successfully.

Client information

Client Information

Automated GitHub Actions environment.

Login information

GitHub Actions using a GitHub App token and GCP Workload Identity Federation.

Anything else we need to know?

The workflows are also encountering an "untrusted directory" error due to changes in how run-gemini-cli evaluates workspaces in headless mode.

Metadata

Metadata

Assignees

Labels

area/platformIssues related to Build infra, Release mgmt, Testing, Eval infra, Capacity, Quota mgmtkind/bugstatus/need-triageIssues that need to be triaged by the triage automation.

Type

No fields configured for Bug.

Projects

Status

Closed

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions