Update prom/prometheus Docker tag to v3.13.1

[renovate]

This PR contains the following updates:

Package	Update	Change
prom/prometheus	minor	v3.12.0 → v3.13.1

---

Release Notes

prometheus/prometheus (prom/prometheus)

v3.13.1: 3.13.1 / 2026-07-10

Compare Source

This is a bugfix release for 3.13 LTS.

• [BUGFIX] TSDB: Fix the head-chunk cache returning samples from the wrong chunk, or spurious not-found errors, to range queries after head-chunk truncation. #​19134

v3.13.0: 3.13.0 / 2026-07-01

Compare Source

This is a Long Term Support LTS release.

• [SECURITY] UI: Bump sanitize-html to fix a cross-site scripting vulnerability (CVE-2026-44990). #​18697
• [CHANGE] UI: Third-party npm dependency licenses are now embedded in the Prometheus binary and served at /assets/third-party-licenses.txt, replacing the npm_licenses.tar.bz2 archive previously shipped in release tarballs and container images. #​18997
• [CHANGE] API: Use SHA-256 instead of SHA-1 to generate rule group pagination tokens. #​18927
• [CHANGE] HTTP clients: Credentials (Authorization header, basic auth, bearer token, OAuth2, configured headers) are no longer forwarded when following a redirect to a different host; affects scraping, remote read/write, alerting, and service discovery. Via prometheus/common v0.69.0 (CVE-2025-4673 CVE-2023-45289). #​18949
• [CHANGE] promtool: Relative file paths in the file passed to --http.config.file are now resolved relative to that config file's directory instead of its parent directory. Via prometheus/common v0.69.0. #​18949
• [CHANGE] PromQL: Rename the min() and max() duration-expression functions (experimental feature flag experimental-duration-expr) to min_of() and max_of() to avoid confusion with the min and max aggregate operators. #​18687
• [FEATURE] API: Add experimental search endpoints to search metric names, label names, and label values. #​18573
• [FEATURE] Discovery/AWS: Add ability to filter RDS instances. #​18859
• [FEATURE] PromQL: Add min_of(a, b) and max_of(a, b) scalar experimental functions, returning the smaller or larger of two scalar values. #​18687
• [FEATURE] PromQL: Add support for smoothed/anchored rate with native histograms. #​18564
• [FEATURE] PromQL: Expose per-query samplesRead (and samplesReadPerStep with stats=all and the promql-per-step-stats feature flag) in the query stats response, and add the prometheus_engine_query_samples_read_total engine counter. samplesRead reflects storage I/O distinct from totalQueryableSamples, which counts samples loaded into the evaluator (and so over-counts when a sample is reused across multiple range-vector windows). #​18081
• [FEATURE] Scrape: Add __convert_classic_histograms_to_nhcb__ internal label to allow per-target override of convert_classic_histograms_to_nhcb scrape configuration via relabeling. #​18840
• [FEATURE] TSDB: Add storage.tsdb.chunk_encoding.floats configuration field to select float chunk encoding (xor or xor2) at runtime, independently of the --enable-feature=xor2-encoding flag. #​18769
• [FEATURE] remote_write: Add Certificate support for ingesting data into an Azure Monitor Workspace. #​18217
• [FEATURE] Scrape: Add __always_scrape_classic_histograms__ and __scrape_native_histograms__ internal labels to allow per-target override of the always_scrape_classic_histograms and scrape_native_histograms scrape configuration via relabeling. #​18929
• [ENHANCEMENT] Release: Container images are now also published to the GitHub Container Registry (ghcr.io). #​18791
• [ENHANCEMENT] PromQL: Prettify fill_left(x) fill_right(x) as fill(x) when both fill values are equal. #​18851
• [ENHANCEMENT] UI: Improve autocompletion after closing a function bracket. #​18894
• [PERF] Labels: Add case-insensitive prefix matching to speed up evaluation of long case-insensitive regular expressions (up to ~2x faster). #​18540
• [PERF] TSDB: Reduce per-sample overhead in chunk population, speeding up affected queries by ~12-15% in benchmarks. #​18699
• [PERF] TSDB: Eliminate unnecessary heap allocations in the V2 histogram WAL decoder, reducing allocations by up to 50% and memory by up to 10% for deployments using native histograms with created-timestamp storage enabled (--enable-feature=created-timestamp-zero-ingestion). #​18813
• [BUGFIX] Discovery/AWS: Fix failure when processing an AWS RDS cluster without instances. #​18845
• [BUGFIX] Fix race condition in initTime that could cause ErrOutOfBounds. #​18629
• [BUGFIX] PromQL: A range query whose end was not aligned to step caused subqueries inside it to evaluate past the parent's last actual step, inflating peakSamples in the query stats and against the query.max-samples limit, and wasting storage I/O reading samples that were never used in the result. #​18081
• [BUGFIX] PromQL: A range query containing an at-modifier-unsafe function over a range-vector with an @ modifier (e.g. predict_linear(metric[60s] @​ T, X)) silently under-counted totalQueryableSamples for steps after step 0. #​18081
• [BUGFIX] PromQL: Fix fill_left/fill_right producing missing samples in range queries when using group_left/group_right. #​18850
• [BUGFIX] PromQL: Fix for resets() and changes() in anchored range extenders with histograms. #​18906
• [BUGFIX] PromQL: Fix panic on 1[5m] smoothed and similar expressions when extended range selectors are enabled. #​18764
• [BUGFIX] PromQL: Fix panic when a smoothed instant vector selector produces no samples for a series. #​18943
• [BUGFIX] PromQL: Fix panic when using a parenthesised plain number as an offset (e.g. foo offset -(5)). #​18768
• [BUGFIX] promtool: Fix panic when parsing exposition text containing empty braces {}. Via prometheus/common v0.69.0. #​18949
• [BUGFIX] Promtool: Fix check healthy and check ready when --url ends with a trailing slash. #​18854
• [BUGFIX] Rules: Close PromQL query after each rule evaluation to ensure resources are released. #​18733
• [BUGFIX] Scaleway SD: Resolve VPC/IPAM-only instances that have no legacy private_ip or public_ip field, but do have private NICs attached. #​18772
• [BUGFIX] TSDB: Do not leak head series when an integer histogram append is rejected (e.g. out-of-order). #​18838
• [BUGFIX] UI: Escape label values offered by PromQL autocomplete. #​18658
• [BUGFIX] TSDB: Fix chunk snapshot encoding for EncXOR2 chunks, preventing corruption on TSDB restart when EncXOR2-encoded series were present. #​18739
• [BUGFIX] TSDB: Store a millisecond timestamp (not a WAL segment number) in walExpiries when a series is evicted via CompactStaleHead/CompactSelectedSeries, so the series's label record is correctly retained in the next WAL checkpoint and replays cleanly. #​18847
• [BUGFIX] TSDB: Prevent loss of samples at the chunk-range boundary when CompactSelectedSeries (and CompactStaleHead) evict the series — the per-slice compaction loop now runs one more iteration so the boundary timestamp is captured in a block before the in-memory copy is removed. #​18849

---

Configuration

📅 Schedule: (in timezone America/New_York)

• Branch creation
  • "after 12am and before 2am on monday"
• Automerge
  • At any time (no schedule defined)

🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.

♻ 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.

[github-actions]

prom/prometheus (docker) v3.12.0 -> v3.13.1

Risk: 🔴 Risk

The Deep Dive

Update Scope

The deployed prom/prometheus container moves from v3.12.0 to the pinned v3.13.1 digest; this is a direct image update in the Prometheus Deployment, with the configuration, persistent volume, Alertmanager, and all other containers unchanged. The target is the first bug-fix release of the 3.13 LTS line.

Performance & Stability

• Range-query correctness after head truncation: v3.13.1 fixes the head-chunk cache, which could return samples from the wrong chunk or false not-found errors after truncation. This applies automatically to this deployment's persistent standalone TSDB and range-query consumers.
• Lower query cost: chunk population is about 12–15% faster in upstream benchmarks, and case-insensitive prefix matching can make long case-insensitive regexes up to about twice as fast; both are automatic when those workloads occur.
• Native-histogram memory path: V2 histogram WAL decoding reduces allocations by up to 50% and memory by up to 10%, but the optimized path additionally requires created-timestamp-zero-ingestion, which is not enabled in the checked Deployment.
• Rule evaluation cleanup: each rule evaluation now closes its PromQL query, automatically releasing resources for the deployment's configured recording and alerting rules.

Features & UX

• Search and query observability APIs: experimental endpoints can search metric names, label names, and label values, but require the absent search-api feature flag and are therefore inactive in the checked Deployment. If enabled later, the Ingress excludes /api/.* from Authentik, so their exposure would need deliberate consideration. Query responses also gain storage-I/O samplesRead statistics and a new engine counter automatically when statistics are requested, while per-step statistics still require the absent promql-per-step-stats flag.
• Per-target histogram controls: relabeling can now override native/classic histogram scraping per target and classic-to-native conversion per target. This deployment globally enables native-histogram scraping, but does not configure the new internal labels, so adopting them is optional.
• PromQL and TSDB opt-ins: smoothed/anchored rates now support native histograms, experimental scalar min_of/max_of functions are added, and float chunks can select XOR or XOR2 in configuration. The required promql-extended-range-selectors, promql-duration-expr, and xor2-encoding feature flags are absent from the checked Deployment, so these remain available but inactive.
• Unused integration additions: AWS RDS discovery gains instance filters and Azure Monitor remote write gains certificate authentication; neither AWS discovery nor remote write exists in the checked Prometheus configuration.
• Operator/UI additions: Images are now also published on GHCR, although this PR keeps Docker Hub, and the UI gains better function-completion behavior.

Security

• CVE-2026-44990, CVSS 9.3: The deployed v3.12.0 lock file contains the advisory's exact vulnerable sanitize-html 2.17.3, while v3.13.1 contains patched 2.17.5; this PR therefore resolves the critical xmp sanitizer-bypass XSS rather than merely avoiding a vulnerable intermediate release. The Prometheus UI is exposed through TLS and Authentik, so the fix is relevant and automatic.
• CVE-2025-4673, CVSS 6.8: Go net/http could retain proxy authentication headers across cross-origin redirects before Go 1.23.10/1.24.4. Both Prometheus versions are built with Go 1.26, so neither v3.12.0 nor v3.13.1 is affected; the broader prometheus/common cross-host stripping is defense-in-depth for Prometheus-configured credentials, not this PR's remediation of the CVE.
• CVE-2023-45289, CVSS 4.3: Go net/http could forward sensitive headers to an incorrectly matched redirect domain before Go 1.21.8/1.22.1. The Go 1.26 builds of both old and proposed Prometheus are outside that range, so this CVE is contextual rather than resolved by the PR. The new Prometheus-level behavior still matters operationally because the checked configuration uses basic-auth and bearer-token scrapes plus Alertmanager.

Key Fixes

• TSDB durability: rejected integer-histogram appends no longer leak head series, WAL checkpoint retention uses the correct timestamp, and selected/stale-head compaction no longer loses boundary samples. These are automatic safeguards for the persistent 100 GiB TSDB.
• PromQL correctness: unaligned range queries no longer over-read and inflate peakSamples, and 3.13 fixes specific panics or wrong results involving extended selectors, empty smoothed selectors, parenthesized negative offsets, fill modifiers, and anchored histogram resets/changes. Searches of the checked local rules found none of those specialized expressions, so they are defensive rather than current-rule fixes.
• Discovery and tooling: AWS RDS cluster handling, Scaleway VPC/IPAM instances, and promtool parsing/URL handling are fixed, but those discovery providers and promtool workflows are absent from the checked deployment configuration.
• UI safety: autocomplete now escapes label values, applying automatically to the deployed UI.

Hazards & Risks

• Authenticated cross-host redirects are an unresolved compatibility boundary: credentials are newly stripped when a redirect changes host. The repository actively configures basic-auth and bearer-token scrapes plus HTTPS Alertmanager, so the changed path is in actual use. Direct hostnames in the file do not prove that those servers never redirect, and CI cannot reach them; any integration that relies on a credentialed cross-host redirect will fail after this update until its target URL or authentication arrangement is corrected. See Further Follow-up for the required pre-merge determination.
• Rule pagination tokens change: rule-group pagination tokens now use SHA-256 rather than SHA-1. No repo-owned pagination client was found; only an external client that persists a token across the rollout would need to restart pagination.
• Experimental duration functions renamed: min() and max() become min_of() and max_of() under promql-duration-expr. The checked Deployment has no feature flags, and local rules do not use these functions, so the breaking experimental rename is inactive here.
• Packaging-only license relocation: third-party license text moves from a container archive to /assets/third-party-licenses.txt. The repository has no consumer of the removed archive.
• XOR2 restart corruption fix: 3.13 repairs chunk snapshots containing EncXOR2 data. XOR2 is not enabled in the checked Deployment, so no migration or recovery prerequisite is introduced.

Further Follow-up

• Quantify authenticated redirect chains before merge: CI inspection established active basic-auth jobs (blackbox, qbittorrent-api-pod, and qbittorrent-api-ingress), bearer-token jobs (hass, Kubernetes API/node/cAdvisor, MetalLB, and karakeep), and HTTPS Alertmanager in the checked configuration, but it cannot contact those private destinations. From an authorized cluster environment, enumerate every active target through /api/v1/targets, include the configured Alertmanager endpoint, and reproduce each authenticated request with its mounted credential while recording the full redirect chain without printing secret headers—for example, use curl --silent --show-error --location --dump-header <protected-file> --output /dev/null <scrape-or-alertmanager-URL> from a protected administrative shell, then compare the authority (scheme://host:port) of every Location hop and securely delete the header capture. Confirmed same-host/no-redirect chains support changing the verdict to renovate:safe; any relied-upon cross-host hop requires changing the configured URL to the final host or arranging authentication at that host before merge, while an unremediated dependency keeps the update incompatible.

Sources

• Prometheus v3.13.0 release
• Prometheus v3.13.1 release
• Prometheus v3.12.0 to v3.13.1 comparison
• Prometheus repository security advisories
• sanitize-html CVE-2026-44990 advisory
• CVE-2025-4673 GitHub advisory
• CVE-2025-4673 Go vulnerability record
• CVE-2023-45289 GitHub advisory
• CVE-2023-45289 Go vulnerability record
• prometheus/common v0.69.0 release

---

🔴 Verdict: Risk

Use renovate:risk until authenticated redirect compatibility is established: v3.13.1 fixes the deployed UI's critical XSS and several TSDB/rule issues, but it also changes authentication behavior on actively used clients, and CI cannot determine whether any private scrape or Alertmanager endpoint relies on cross-host redirects. Confirming no such redirect supports renovate:safe; any relied-upon cross-host redirect must be remediated before merge.