Problem
When a container build fails, the build output (stdout/stderr) stored in the log entries and GitHub check run annotations is truncated:
out: [1/4] STEP 1/6: FROM node:22-slim AS frontend-builder... [truncated 11181 chars]
error: npm notice... [truncated 2013 chars]
This makes it impossible to diagnose build failures from the logs alone. The actual error (e.g., npm error code EINTEGRITY) is buried in the truncated portion.
Expected Behavior
The full build output should be stored in the log entries. The GitHub check run annotation can remain truncated (GitHub has API limits), but the webhook server's own logs should preserve the complete output for debugging via MCP tools (get_log_entries, get_step_logs).
Current Behavior
Both the log entries and the check run text are truncated at the same length, making it impossible to find the actual build error through any channel.
Suggested Fix
- Store the full build output in the webhook server's log entries (no truncation)
- Only truncate when writing to GitHub check run annotations (which have API size limits)
Done
Problem
When a container build fails, the build output (stdout/stderr) stored in the log entries and GitHub check run annotations is truncated:
This makes it impossible to diagnose build failures from the logs alone. The actual error (e.g.,
npm error code EINTEGRITY) is buried in the truncated portion.Expected Behavior
The full build output should be stored in the log entries. The GitHub check run annotation can remain truncated (GitHub has API limits), but the webhook server's own logs should preserve the complete output for debugging via MCP tools (
get_log_entries,get_step_logs).Current Behavior
Both the log entries and the check run text are truncated at the same length, making it impossible to find the actual build error through any channel.
Suggested Fix
Done