feat(container): update image victoriametrics/victoria-metrics to v1.113.0

[trap-renovate]

This PR contains the following updates:

Package	Update	Change
victoriametrics/victoria-metrics	minor	v1.112.0 -> v1.113.0

---

Release Notes

VictoriaMetrics/VictoriaMetrics (victoriametrics/victoria-metrics)

v1.113.0

Compare Source

v1.113.0

Released at 2025-03-07

Update note 1: vmsingle and vmagent include a fix which enforces IPv6 addresses escaping for containers discovered with Kubernetes service-discovery and role: pod which do not have exposed ports defined. This means that address for these containers will always be wrapped in square brackets, this might affect some relabeling rules which were relying on previous behaviour.

Update note 2: vmalert disallows using time buckets stats pipe in alerting or recording rules with VictoriaLogs as datasource. Time buckets used with stats query API may produce unexpected results for user and result into cardinality issues.

Update note 3: vmalert disallows specifying eval_offset and eval_delay options in the same group. The eval_offset option ensures the group is evaluated at the exact offset in the range of [0...interval]. However, with eval_delay, this behavior cannot be guaranteed without further adjusting the evaluation time, which could lead to more confusion.

• FEATURE: upgrade Go builder from Go1.23.6 to Go1.24. See Go1.24 release notes.

• FEATURE: provide alternative registry for all VictoriaMetrics components at Quay.io.

• FEATURE: vmsingle and vmstorage in VictoriaMetrics cluster: add a new flag --storage.trackMetricNamesStats and a new HTTP API - /api/v1/status/metric_names_stats. It allows to track how frequent ingested metric names are used during querying. See this issue for details and related docs

• FEATURE: data ingestion: make KeyValueList, ArrayValue OpenTelemetry protocol for metrics attributes label values compatible with open-telemetry-collector format. See this issue.

• FEATURE: vmalert: disallow using time buckets stats pipe in VictoriaLogs rule expressions. Such construction produces meaningless results for stats query API and may lead to cardinality issues.

• FEATURE: vmalert: remove random sleep before a group starts when eval_offset is specified, because eval_offset already disperses the group evaluation time, serving the same purpose as the random sleep. This change also enables chaining groups, see this doc and this issue.

• FEATURE: vmalert-tool: add command-line flag -httpListenPort to specify the port used during testing. If not provided, a random unoccupied port will be assigned. See this issue.

• FEATURE: vmalert-tool: make the temporary storage path for unittest unique, allowing user to run multiple vmalert-tool processes on a single host simultaneously. See this issue.

• FEATURE: alerts: add alerting rule TooHighQueryLoad to notify user when VictoriaMetrics or vmselect weren't able to serve requests in timely manner during last 15min.

• FEATURE: dashboards/single and dashboards/cluster: add panel Deduplication rate that shows how many samples are deduplicated during merges or read queries by VictoriaMetrics components.

• FEATURE: dashboards/single and dashboards/cluster: add panel Number of snapshots that shows the max number of snapshots across vmstorage nodes. This panel should help in disk usage troubleshooting.

• FEATURE: dashboards/single and dashboards/cluster: account for samples dropped according to relabeling config in Samples dropped for last 1h panel.

• FEATURE: dashboards/single and dashboards/cluster: show number of parts in the last partition on LSM parts max by type panel. Before, the resulting graph could be skewed by the max number of parts across all partitions. Displaying parts for the latest partition is the correct way to show if storage is currently impacted by merge delays.

• FEATURE: dashboards/cluster: add panel Partial query results that shows the number of served partial responses by vmselects.

• BUGFIX: Single-node VictoriaMetrics and vmstorage: fix the incorrect caching of extMetricsIDs when a query timeout error occurs. This can lead to incorrect query results. Thanks to @​changshun-shi for the bug report issue.

• BUGFIX: vmctl: respect time filter when exploring time series for influxdb mode. See this issue for details.

• BUGFIX: vmsingle: properly apply global relabeling configuration, defined with -relabelConfig flag, for metrics scrapped with -promscrape.config. Bug was introduces in v1.108.0. See this issue.

• BUGFIX: vmbackupmanager: properly propagate an error message when applying retention policy fails. Previously, an actual error messages was discarded.

• BUGFIX: vmgateway: fix data query in rate limiter. The bug was introduced in this commit starting from v1.106.0.

• BUGFIX: vmgateway: properly apply the rate limiter for the rows_inserted limit type. Previously, the rate limit for this type was ignored.

• BUGFIX: vmgateway: properly handle HTTP requests with path ending with a trailing / when using the rate limiter. Previously, the trailing slash was removed and caused an incorrect redirect path when visiting VMUI. Thanks to @​jindov for the bug report issue.

• BUGFIX: vmsingle and vmagent: properly escape IPv6 address in Kubernetes service-discovery with role: pod for containers without exposed ports. See this issue.

• BUGFIX: vmalert-tool: clean up the temporary storage path when process is terminated by SIGTERM or SIGINT. Previously, unclean shut down might affect the next run.

• BUGFIX: vmui: fix an infinite loader on the Downsampling filters debug page when provided configuration matches no series. See this issue.

• BUGFIX: MetricsQL: fix filters pushdown logic for expression like foo{a="a"} ifnot bar{a="b"}. Previously, filters from right operand were incorrectly propagated to the left operand and could result in empty query results even if foo{a="a"} matches time series. See this issue.

• BUGFIX: vmsingle, vmselect in VictoriaMetrics cluster: prevent possible panic for foo @​ bar expression when first sample in bar starts with NaN or starts long after first sample in foo. Now, VM will try to find first non-NaN value in bar and could yield an error @ modifier must return a non-NaN value if it won't find it. See this issue.

• BUGFIX: Single-node VictoriaMetrics and vmstorage: prevent panic when using with rules that have zero interval: -downsampling.period=5m:5m,0s:0s. Such rule configuration shouldn't be rejected and cause an error when used. See this issue.

---

Configuration

📅 Schedule: Branch creation - At any time (no schedule defined), Automerge - At any time (no schedule defined).

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

♻ Rebasing: Whenever PR becomes conflicted, 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 has been generated by Renovate Bot.

[github-actions]

Diff

@@ spec.template.spec.containers.victoria-metrics.image @@
# apps/v1/StatefulSet/monitor/victoria-metrics
! ± value change
- victoriametrics/victoria-metrics:v1.112.0
+ victoriametrics/victoria-metrics:v1.113.0