[SECURITY] Update Rust crate opentelemetry_sdk to 0.33.0 [SECURITY]

[renovate]

This PR contains the following updates:

Package	Type	Update	Change
opentelemetry_sdk (source)	dependencies	minor	0.32.0 → 0.33.0

---

opentelemetry_sdk has unbounded memory allocation in W3C Baggage propagation

CVE-2026-48504 / GHSA-w9wp-h8wv-79jx

More information

Details

Summary

BaggagePropagator::extract_with_context in opentelemetry_sdk did not enforce the W3C Baggage size limits before parsing an inbound baggage header. A large attacker-controlled header could cause unnecessary CPU work and short-lived heap allocations while parsing entries that would later be discarded by the SDK's baggage storage limits.

The SDK now applies limits aligned with the W3C Baggage limits:

• 64 list-members
• 8192 bytes total

Impact

Services that accept untrusted inbound propagation headers may experience increased per-request resource usage when processing oversized baggage headers. This can contribute to denial-of-service risk, especially when application or transport-level header limits are absent or configured above the W3C Baggage limits.

The impact is limited to availability. This issue does not expose telemetry data, modify telemetry data, or allow code execution.

Patches

Upgrade opentelemetry_sdk to version 0.32.1 or later.

Version 0.32.1 rejects baggage header values larger than 8192 bytes and limits extraction to the first 64 list-members.

Workarounds

If upgrading immediately is not possible, reject or limit inbound baggage headers larger than 8192 bytes before invoking OpenTelemetry propagation extraction. This can be enforced at a proxy, gateway, middleware layer, or custom carrier boundary.

Resources

• W3C Baggage limits: https://www.w3.org/TR/baggage/#limits
• Related OpenTelemetry Java advisory: GHSA-rcgg-9c38-7xpx
• Related OpenTelemetry Go advisory: GHSA-mh2q-q3fh-2475
• CVE-2026-48504

Credit

tonghuaroot

Severity

• CVSS Score: 5.3 / 10 (Medium)
• Vector String: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L

References

• https://github.com/open-telemetry/opentelemetry-rust/security/advisories/GHSA-w9wp-h8wv-79jx
• https://github.com/advisories/GHSA-w9wp-h8wv-79jx

This data is provided by the GitHub Advisory Database (CC-BY 4.0).

---

Release Notes

open-telemetry/opentelemetry-rust (opentelemetry_sdk)

v0.33.0

Released 2026-Sep-18

• Publicly export the OTEL_*/OTEL_*_DEFAULT environment variable name and
  default value constants for BatchSpanProcessor (opentelemetry_sdk::trace),
  BatchLogProcessor (opentelemetry_sdk::logs), and PeriodicReader
  (opentelemetry_sdk::metrics), so downstream configuration systems can read
  the SDK's spec-defined defaults programmatically instead of duplicating
  them. As part of this, PeriodicReader's previously-private
  DEFAULT_INTERVAL/METRIC_EXPORT_INTERVAL_NAME constants were renamed
  to OTEL_METRIC_EXPORT_INTERVAL_DEFAULT/OTEL_METRIC_EXPORT_INTERVAL to
  match the naming convention already used elsewhere.
  (#​3623)
• Added SDK self-observability metrics, feature-gated behind
  experimental_metrics_bound_instruments: otel.sdk.log.created counts log
  records submitted to the SDK; otel.sdk.processor.log.processed and
  otel.sdk.processor.span.processed count records and spans submitted to an
  exporter by batch and simple processors, with error.type reporting items
  dropped before submission; and otel.sdk.processor.log.queue.capacity
  reports the configured BatchLogProcessor queue capacity.
  (#​3514,
  #​3608,
  #​3609,
  #​3611)
• Made futures-channel, futures-executor, futures-util, and thiserror
  optional, enabling a minimal SDK build. With default-features = false, the
  SDK's only dependency is the opentelemetry API crate.
  (#​3593)
• Bound instruments are now available for Gauge and UpDownCounter via the
  new BoundGauge<T> and BoundUpDownCounter<T> types exposed by the
  opentelemetry crate. Requires the experimental_metrics_bound_instruments
  feature.
• Fixed a race in BatchSpanProcessor and BatchLogProcessor where a
  span/log enqueued just before force_flush() or shutdown() could be
  missed by the flush and dropped at shutdown: the pending-item counter is
  now incremented before enqueueing (and reverted if the queue is full), so
  the worker's counter snapshot can no longer under-count items already in
  the queue (#​3453).
• Default SDK Resource construction now falls back to unknown_service under
  Miri instead of calling std::env::current_exe(), avoiding an abort in Miri
  isolation mode while preserving the normal
  unknown_service:<process.executable.name> fallback outside Miri.
• Fixed asynchronous counters (ObservableCounter, ObservableUpDownCounter)
  using delta temporality reporting incorrect deltas when observed attributes
  were recorded in an unsorted key order.

---

Configuration

📅 Schedule: (UTC)

• Branch creation
  • At any time (no schedule defined)
• Automerge
  • At any time (no schedule defined)

🚦 Automerge: Enabled.

♻ Rebasing: Whenever PR is behind base branch, or you tick the rebase/retry checkbox.

🔕 Ignore: Close this PR and you won't be reminded about this update again.

---

• If you want to rebase/retry this PR, check this box

---

This PR was generated by Mend Renovate. View the repository job log.