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
{{ message }}
Repository navigation
Reduce HIGH findings on \otel/opentelemetry-collector-contrib\ pinned digest (4 HIGH, 0 CRITICAL — 2026 Go stdlib/x-net wave) #115
Reduce HIGH findings on otel/opentelemetry-collector-contrib (4 HIGH, 0 CRITICAL)
Status
Open. Surfaced 2026-10-05..10 by the weekly scan after new Go CVEs were published against existing package versions (DB movement against a frozen image — the exact scenario flagged in docs/container_security_scan.md's re-scan notice). The collector itself did not drift; nothing broke.
Pin
otel/opentelemetry-collector-contrib@sha256:fd328de2552466ad78385e1b1289c3f2402b1c45f265b252aab1955b42845ac1 (0.161.0; 0 HIGH/0 CRIT at pin-time 2026-10-03).
Findings (Trivy 0.74.0, DB 2026-10-10, HIGH+CRITICAL, ignore-unfixed)
Registry :latest = sha256:8b9211db... scanned on the same DB: identical 4 findings. Upstream collector releases are not yet built with fixed Go/x-net, so NO re-pin was adopted (gate: adopt only on verified equal-or-better). Same gate correctly rejected the newer grype:debug (3518aea7..., same 4 findings) in favor of our clean 0-finding 1d10432f... deployment pin.
Why NOT in .trivyignore
Core default-profile path (the pipeline's ingest spine). These IDs are shared with #83/#84 images, and ID-global suppression would also blind future OTel regressions. Keep visible until rebuilt.
Resolution path
The CI matrix entry now floats to :latest as the drift monitor (this issue's companion PR; same doctrine as clamav/#111 and goflow2/dcgm). The weekly/push scan job Trivy scan (otel-collector-contrib) dropping to 0 on floating :latest = the signal to verify + re-pin compose to the patched digest.
Acceptance criteria
Compose re-pinned to a verified 0-HIGH/0-CRITICAL contrib digest built with Go >= 1.26.9/1.27.2 and x/net >= 0.60.0, docs rows + this issue closed; or
Upstream declares the stdlib DoS vectors non-exploitable in the collector (document + VEX instead of re-pin).
Owner
TBD. Active in the default compose path (highest-priority image in the stack).
Monitor restored + re-pin gate result (2026-10-10)
Monitor. PR #116 (squash f29a2a3, merged 2026-10-11 02:23 UTC) floated the CI matrix otel-collector-contrib entry so it tracks the upstream rebuild. First post-merge run on main (run 38105001331) passed all 12 scan jobs; Trivy scan (otel-collector-contrib) reporting 4 HIGH is now expected steady state (gate is CRITICAL-only). When upstream rebuilds with fixed Go (>= 1.26.9 / 1.27.2) and/or golang.org/x/net 0.60.0, findings drop to 0 — that drop is the re-pin trigger.
Gate result. Registry :latest = sha256:8b9211db… scanned with current Trivy DB yields the identical 4 findings as the deployed pin sha256:fd328de2… (CVE-2026-78667, CVE-2026-78669 stdlib+x/net, CVE-2026-97031). Lateral move → no re-pin; compose keeps fd328de2….
Suppression. The new-wave IDs are deliberately not added to .trivyignore: ID-global suppression would blind this issue's own regression signal on a fixable, core-path image. Revisit only after the pin moves to a >= fixed build.
Related: the 19 new-wave alerts (#495–#513) stay open as current upstream-pending signals; umbrella tracking on #82.
Activity
fix(ci): float otel-collector-contrib matrix entry as #115 rebuild mo…
JJediny commentedon Oct 11, 2026
Monitor restored + re-pin gate result (2026-10-10)
Monitor. PR #116 (squash
f29a2a3, merged 2026-10-11 02:23 UTC) floated the CI matrixotel-collector-contribentry so it tracks the upstream rebuild. First post-merge run onmain(run38105001331) passed all 12 scan jobs;Trivy scan (otel-collector-contrib)reporting 4 HIGH is now expected steady state (gate is CRITICAL-only). When upstream rebuilds with fixed Go (>= 1.26.9 / 1.27.2) and/orgolang.org/x/net0.60.0, findings drop to 0 — that drop is the re-pin trigger.Gate result. Registry
:latest=sha256:8b9211db…scanned with current Trivy DB yields the identical 4 findings as the deployed pinsha256:fd328de2…(CVE-2026-78667, CVE-2026-78669 stdlib+x/net, CVE-2026-97031). Lateral move → no re-pin; compose keepsfd328de2….Suppression. The new-wave IDs are deliberately not added to
.trivyignore: ID-global suppression would blind this issue's own regression signal on a fixable, core-path image. Revisit only after the pin moves to a >= fixed build.Related: the 19 new-wave alerts (#495–#513) stay open as current upstream-pending signals; umbrella tracking on #82.