[release/17.0] Retry transient GraphQL failures in milestone changelog fetch-data

.github/workflows/README.md

@@ -90,6 +90,23 @@ Node 20 action. There is no supported Node 24 upgrade for that dependency;
 replacing it requires a separate migration of its authentication and locking
 behavior, rather than merely changing an action pin.
 
+## Milestone changelog fetch diagnostics
+
+`milestone-changelog.md` retries PR-list requests up to four times, waiting
+15, 30, and 60 seconds between attempts. HTTP 5xx errors and the
+`unexpected end of JSON input` diagnostic are retryable. When `gh` succeeds
+but returns invalid or empty JSON, validation also retries, even if stderr
+contains warnings. Other command failures and valid JSON of the wrong type
+fail immediately.
+
+Each failed attempt logs the output filename, command and validation exit codes,
+stderr from both tools, and the first 4096 bytes of the response with its total
+byte count. Response previews are bounded because PR bodies can be large.
+Diagnostic lines are prefixed and legacy `##[` command markers are escaped so
+response text is not interpreted as GitHub Actions workflow commands.
+Successful requests also replay any `gh` warnings;
+their complete JSON is kept in the normal data files rather than printed.
+
 ## Auto-sec dependency reconciliation
 
 `auto-sec.md` runs every 12 hours (and on demand). It reconciles open Dependabot

.github/workflows/milestone-changelog.md

@@ -74,6 +74,68 @@ jobs:
           DATA_DIR="/tmp/gh-aw/pr-data"
           mkdir -p "$DATA_DIR"
 
+          # Run `gh pr list` with bounded retry. The large (--limit 5000)
+          # GraphQL queries below occasionally hit a transient 502/504 from
+          # GitHub's GraphQL gateway or get a truncated response body. Retry
+          # only those transient failures (4 attempts, 15s/30s/60s backoff);
+          # Invalid JSON after a successful command is retryable even if gh
+          # printed warnings. Other command errors and valid non-array JSON
+          # fail immediately. Usage: gh_pr_list_with_retry OUT ARGS...
+          gh_pr_list_with_retry() {
+            local out="$1"; shift
+            local delays=(15 30 60)
+            local attempt=1
+            local diagnostics err validation_err
+            diagnostics=$(mktemp -d)
+            err="$diagnostics/gh.stderr"
+            validation_err="$diagnostics/jq.stderr"
+            while true; do
+              local status=0 validation_status=0
+              local transient=false
+              : > "$validation_err"
+              if gh pr list "$@" > "$out" 2> "$err"; then
+                if jq -s -e 'if length == 1 and (.[0] | type == "array") then true else error("Expected exactly one JSON array") end' \
+                    "$out" > /dev/null 2> "$validation_err"; then
+                  tr '\r' '\n' < "$err" | sed 's/##\[/##\\[/g; s/^/gh stderr: /' >&2
+                  rm -f "$err" "$validation_err"
+                  rmdir "$diagnostics"
+                  return 0
+                else
+                  validation_status=$?
+                fi
+                if ! jq -s -e 'length > 0' "$out" > /dev/null 2>&1; then
+                  transient=true
+                fi
+              else
+                status=$?
+                # Examples: "HTTP 502: Bad Gateway", "unexpected end of JSON input".
+                if grep -qE 'HTTP 5[0-9][0-9]|unexpected end of JSON input' "$err"; then
+                  transient=true
+                fi
+              fi
+
+              echo "gh pr list attempt $attempt failed for $out (gh exit=$status, jq exit=$validation_status)" >&2
+              tr '\r' '\n' < "$err" | sed 's/##\[/##\\[/g; s/^/gh stderr: /' >&2
+              tr '\r' '\n' < "$validation_err" | sed 's/##\[/##\\[/g; s/^/jq stderr: /' >&2
+              # PR bodies can be large and contain workflow-command syntax.
+              # Bound the preview and prefix every line, including CR-delimited lines.
+              # Legacy "##[command]" markers are recognized even after a prefix.
+              echo "Response preview (first 4096 bytes of $(wc -c < "$out") bytes):" >&2
+              head -c 4096 "$out" | tr '\r' '\n' | sed 's/##\[/##\\[/g; s/^/stdout: /' >&2
+              printf '\n' >&2
+              if [ "$transient" != true ] || [ "$attempt" -gt "${#delays[@]}" ]; then
+                rm -f "$err" "$validation_err"
+                rmdir "$diagnostics"
+                echo "::error::gh pr list failed after $attempt attempt(s)"
+                return 1
+              fi
+              local delay="${delays[$((attempt - 1))]}"
+              echo "::warning::gh pr list attempt $attempt failed with a transient error; retrying in ${delay}s"
+              sleep "$delay"
+              attempt=$((attempt + 1))
+            done
+          }
+
           # 0. Clone the memory branch to read existing state (prs/, changes/,
           #    feedback-issue.json). The full content is copied to
           #    $DATA_DIR/memory/$MILESTONE/ so the agent can read it directly.
@@ -144,11 +206,13 @@ jobs:
           rm -f "$CLONE_ERR"
 
           # 1. Fetch ALL merged PRs in milestone, sorted by merge date ascending
-          gh pr list --repo "$REPO" --state merged --limit 5000 \
+          gh_pr_list_with_retry "$DATA_DIR/all-milestone-prs.raw.json" \
+            --repo "$REPO" --state merged --limit 5000 \
             --search "milestone:$MILESTONE" \
-            --json number,title,body,author,mergedBy,mergedAt,labels,additions,deletions,changedFiles \
-            | jq 'sort_by(.mergedAt)' \
+            --json number,title,body,author,mergedBy,mergedAt,labels,additions,deletions,changedFiles
+          jq 'sort_by(.mergedAt)' "$DATA_DIR/all-milestone-prs.raw.json" \
             > "$DATA_DIR/all-milestone-prs.json"
+          rm -f "$DATA_DIR/all-milestone-prs.raw.json"
 
           TOTAL=$(jq length "$DATA_DIR/all-milestone-prs.json")
           echo "Total merged PRs in milestone: $TOTAL"
@@ -226,11 +290,13 @@ jobs:
             DOCS_MERGED_RANGE="merged:${MILESTONE_START}..${BATCH_CUTOFF}"
             echo "Capping docs PRs to merged on or before $BATCH_CUTOFF (batch is full)"
           fi
-          gh pr list --repo "$DOCS_REPO" --state merged --limit 5000 \
+          gh_pr_list_with_retry "$DATA_DIR/all-docs-prs.raw.json" \
+            --repo "$DOCS_REPO" --state merged --limit 5000 \
             --search "$DOCS_MERGED_RANGE" \
-            --json number,title,body,author,mergedBy,mergedAt,labels,additions,deletions,changedFiles,files \
-            | jq 'sort_by(.mergedAt)' \
+            --json number,title,body,author,mergedBy,mergedAt,labels,additions,deletions,changedFiles,files
+          jq 'sort_by(.mergedAt)' "$DATA_DIR/all-docs-prs.raw.json" \
             > "$DATA_DIR/all-docs-prs.json"
+          rm -f "$DATA_DIR/all-docs-prs.raw.json"
 
           DOCS_TOTAL=$(jq length "$DATA_DIR/all-docs-prs.json")
           echo "Total docs PRs merged since $MILESTONE_START: $DOCS_TOTAL"