-
Notifications
You must be signed in to change notification settings - Fork 36
Use info-level logging for missing agent output files #10128
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Merged
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
When agent fails before producing safe-outputs (e.g., due to API errors like "missing finish_reason"), the agent-output artifact doesn't exist. This causes the conclusion job to log error messages when trying to read the non-existent file. Changed load_agent_output.cjs to use core.info instead of core.error when the file cannot be read, making the logging less alarming for this normal failure scenario. Updated corresponding test to match new behavior. This resolves the confusing error messages in the CI Cleaner workflow failure logs. Co-authored-by: mnkiefer <8320933+mnkiefer@users.noreply.github.com>
Copilot
AI
changed the title
[WIP] Fix CI Cleaner workflow failure issue
Use info-level logging for missing agent output files
Jan 15, 2026
Copilot AI
added a commit
that referenced
this pull request
Jan 15, 2026
The loadAgentOutput function was changed to use core.info instead of core.error for missing files (see #10128). This is a normal scenario when the agent fails before producing safe-outputs, so info-level logging is appropriate. Co-authored-by: Mossaka <5447827+Mossaka@users.noreply.github.com>
17 tasks
Mossaka
added a commit
that referenced
this pull request
Jan 15, 2026
…ess (#9999) * Initial plan * Initial investigation of Codex MCP DNS resolution issue Co-authored-by: Mossaka <5447827+Mossaka@users.noreply.github.com> * Add host.docker.internal to CodexDefaultDomains for MCP gateway access Fixes DNS resolution failure when Codex agent tries to connect to MCP Gateway via host.docker.internal from within AWF container. Co-authored-by: Mossaka <5447827+Mossaka@users.noreply.github.com> * Add api.githubcopilot.com to domain allowlist for GitHub MCP remote mode When using `tools.github.mode: remote`, the Codex agent needs to connect to api.githubcopilot.com. This wasn't being added to the firewall allowlist because the URL is set internally, not in the raw frontmatter. - Add GitHubCopilotMCPDomain constant - Detect `mode: remote` in extractHTTPMCPDomains and add the domain - Add tests for GitHub MCP remote/local mode domain extraction Co-authored-by: Mossaka <5447827+Mossaka@users.noreply.github.com> * Regenerate workflows after merging main * Fix DNS resolution by using gateway IP directly - Change MCP gateway domain from host.docker.internal to 172.30.0.1 - This fixes DNS resolution issues in Rust-based codex engine - Docker's embedded DNS doesn't properly resolve host.docker.internal - Using the gateway IP directly avoids DNS resolution altogether * Update MCP Gateway schema to allow IPv4 addresses - Add IPv4 address pattern to domain field in schema - This allows using gateway IP (172.30.0.1) directly - Fixes DNS resolution issues with host.docker.internal * Revert "Update MCP Gateway schema to allow IPv4 addresses" This reverts commit 8ce85d6. * Revert "Fix DNS resolution by using gateway IP directly" This reverts commit 1bdb5f2. * Fix codex DNS resolution by resolving host.docker.internal in converter - Modify convert_gateway_config_codex.sh to resolve host.docker.internal to 172.30.0.1 - This avoids DNS resolution issues in Rust's async resolver - Gateway still sees host.docker.internal (schema-valid) - Codex gets IP directly (DNS-free) * Add --sandbox danger-full-access to disable Codex CLI internal sandbox The Codex CLI has a built-in Landlock sandbox that blocks file writes and shell commands even with --full-auto. When running inside AWF (which already provides container-level sandboxing), we need to disable Codex's internal sandbox to allow file operations and shell commands. This fixes the errors: - File write ❌: sandbox prevented file creation - Bash cat ❌: sandbox prevented shell command Co-authored-by: Mossaka <5447827+Mossaka@users.noreply.github.com> * Fix noop.test.cjs: update test to expect core.info instead of core.error The loadAgentOutput function was changed to use core.info instead of core.error for missing files (see #10128). This is a normal scenario when the agent fails before producing safe-outputs, so info-level logging is appropriate. Co-authored-by: Mossaka <5447827+Mossaka@users.noreply.github.com> * Increase timeout for Go handler tests to allow for compilation - Add 30 second timeout to all Go tests (was 10s default) - Go compilation in CI can be slow - Fixes test timeout failures in CI --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: Mossaka <5447827+Mossaka@users.noreply.github.com> Co-authored-by: Jiaxiao Zhou <duibao55328@gmail.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
When agents fail before producing safe-outputs (e.g., due to API errors like
missing finish_reason), the conclusion job logs##[error]Error reading agent output fileeven though the missing file is expected. This obscures the actual failure cause.Changes
load_agent_output.cjs: Usecore.info()instead ofcore.error()when file doesn't existload_agent_output.test.cjs: Update test expectations to verifycore.infois calledExample
Before (error level):
After (info level):
The actual agent failure (API error) remains visible in agent job logs. Downstream handlers already gracefully handle missing files.
Original prompt
💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.