You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Unify egress allowlist by merging connected and disconnected modes
CI captures egress traffic from both connected and disconnected e2e
deployments. However, both jobs run on the same hypervisor with shared
caching (pip, Ansible collections), causing the disconnected run to
skip traffic already cached by connected. This produces misleading
per-mode splits.
Merge all captured logs into a single unified allowlist grouped only by
caller category (ansible, pip, podman, etc.).
Changes:
- analyze_egress.py: Accept multiple log files, output complete YAML
- update_allowlist.sh: Use find to collect all log files
- Workflows: Consolidate artifact downloads, remove mode arguments
- Test fixtures: Update to unified format
Signed-off-by: Rafa Porres Molina <rporresm@redhat.com>
Assisted-by: Claude Sonnet 4.5 <noreply@anthropic.com>
# ── Compare with current allow list ───────────────────────────────────────────
47
37
48
38
if [ !-f"$ALLOWLIST" ];then
49
39
# First run — treat an empty file as the current state
50
-
printf"# Auto-generated by scripts/egress/analyze_egress.py\n# Do not edit manually — updated by the egress-capture CI workflow\negress:\n connected:\n []\n disconnected:\n []\n">"$ALLOWLIST"
40
+
printf"# Auto-generated by scripts/egress/analyze_egress.py\n# Do not edit manually — updated by the egress-capture CI workflow\negress: {}\n">"$ALLOWLIST"
0 commit comments