Shipped · Databases

What VictoriaMetrics/VictoriaMetrics shipped

The public repository of Victoriametrics · github.com/VictoriaMetrics/VictoriaMetrics

Written by FoxPlug from the repository's own releases; not affiliated with Victoriametrics. Paste your own repository in the box below and FoxPlug writes the same weekly update for your product. This page is an automatic summary of the releases, pull requests and commits in github.com/VictoriaMetrics/VictoriaMetrics; Victoriametrics did not write it and does not use or endorse FoxPlug. Every line links to the change it describes.

Get a weekly update like this for your product, free

Or Use it as a GitHub Action

See the update FoxPlug writes for VictoriaMetrics, free

Week of November 2, 2026

What shipped

Why it matters

This week brings a cluster release plus fixes across vmselect, vmagent, vmstorage and vmalert that reduce false errors, panics and unnecessary connection churn. Users also get clearer control over persistent queue paths, cgroupv1 resource limits and Azure workload identity discovery.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

V1.156.0-cluster is out. This week: vmselect no longer counts canceled queries as http errors and checks sub-query series limits after step alignment; vmalert keeps original labels in staleness markers; vmagent adds per-URL queue paths and Azure workload identity auth.

Post for LinkedIn

VictoriaMetrics v1.156.0-cluster is released. This week's changes: vmselect stops increasing vm_http_request_errors_total for client-canceled queries and checks the sub-query series limit after start/end alignment, avoiding a runtime panic. vmalert now keeps original labels in recording rule staleness markers. vmagent can configure the persistent queue path per remote write URL and adds WorkloadIdentity authentication for Azure service discovery. lib/cgroup now checks systemd-slice limits on cgroupv1 like cgroupv2. vmstorage exposes the working set cache mode via vm_fastcache_mode, and deployment rules no longer alert for intentionally disabled backups.

Week of October 26, 2026

What shipped

Why it matters

This week brought two feature releases and a patch release, including a breaking change to how /admin/tenants behaves in cluster mode when no time range is given, so users of that endpoint should check their queries. Stream aggregation received two correctness fixes for counter resets and flush timing, and average functions now handle gaps properly. The vmauth latency fix addresses slow proxying in token-authenticated setups without configured users.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

V1.155.0 and v1.154.0 are out. /admin/tenants without start and end now returns current-day tenants in cluster mode, a breaking change. v1.154.1 fixes vmauth delays of 2-3s for token requests with no users configured. Also: stream aggregation rate and average fixes, and Kafka pro

Post for LinkedIn

This week VictoriaMetrics shipped v1.155.0, v1.154.0 and the v1.154.1 patch. In v1.155.0, vmselect in cluster mode now returns tenants for the current day when /admin/tenants is called without start and end query args; this is a breaking change, since the endpoint previously returned all tenants. v1.154.1 fixes vmauth slow requests with auth tokens when no users are configured, which were delayed 2-3 seconds before proxying to unauthorized_user. Stream aggregation got two fixes: false counter resets in rate outputs with enable_windows, and a flush delay capped at 90 percent of the interval. running_avg and range_avg now count only non-NaN points, the vmagent dashboard shows Kafka producer latency, and the anomaly detection docs gained a model selection guide and Online MAD decay details.

Week of October 19, 2026

What shipped

Why it matters

This week brought a cluster release and fixes that reduce load and avoid failures in storage searches, tenant queries, and handshakes. Users get more stable vmselect, vminsert, and vmalert behavior, improved Kubernetes discovery, and more predictable vmui and dashboard experiences.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

V1.155.0-cluster is out. This week: per-day search concurrency is capped, /admin/tenants defaults to the current day, handshake timeouts no longer drop metadata, and vmalert restores alert state compatibly with Prometheus. vmui and dashboard fixes also landed.

Post for LinkedIn

VictoriaMetrics v1.155.0-cluster is released. This week's changes include a cap on per-day index search concurrency to avoid goroutine spikes, a current-day default for /admin/tenants to reduce storage load, and a fix so handshake timeouts do not force vminsert into legacy mode. vmalert now restores alert state from remote read in a Prometheus-compatible way, Kubernetes endpoints discovery uses the correct namespace for pods in other namespaces, and cached metricIDs are protected from concurrent mutation. vmui limits tenant discovery to the selected time range, dashboards have consistent resource filters, and a new -healthcheck flag helps with Docker health checks.

Week of October 12, 2026

What shipped

Why it matters

This week brought a new release, v1.154.0, with security-related changes in the Go builder and vmauth. Storage and vmagent also got changes aimed at safer search range handling, lower memory use during Kubernetes target rebuilds, and rejecting malformed native blocks. On the UI and docs side, vmui gained a tenant column and stricter Server URL editing, and new guides cover data recovery and unix socket scraping.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

V1.154.0 is out: Go builder updated and vmauth slows invalid auth requests. Also this week: lower vmagent memory during Kubernetes target rebuilds, tenant column in vmui, and a new data recovery guide.

Post for LinkedIn

V1.154.0 is released, with the Go builder upgraded from Go1.27.1 to Go1.27.2 and vmauth slowing down requests with invalid auth tokens even when unauthorized_user has URLs configured. In storage, search time range adjustment moved into indexDB, and native blocks with reversed timestamp ranges are now rejected rather than risking a panic. vmagent retains less memory during full Kubernetes service-discovery target rebuilds. vmui adds a tenant column to Top Queries and Active Queries, disables Server URL editing by default, and updates alert styling. Docs gained a data recovery guide and a vmagent guide for scraping targets via unix sockets.

Week of October 5, 2026

What shipped

Why it matters

This week brings a cluster release plus storage changes that free idle buffers and skip the global index when the per-day index is enabled, which affects memory use and query behavior. Streaming aggregation results are corrected for enable_windows, and several networking and alerting fixes reduce leaks, races and reload failures.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

This week: v1.154.0-cluster out. Storage frees idle rawRowsShard buffers, skips global index with per-day index, streamaggr fixes enable_windows totals, vmalert fixes relabel race and retries notifier reload. vmanomaly docs updated for v1.31.0 and UI v1.10.0.

Post for LinkedIn

This week VictoriaMetrics shipped v1.154.0-cluster and a set of fixes across storage, streaming aggregation and alerting. lib/storage now releases idle rawRowsShard buffers after periodic flush and does not search the global index when the per-day index is enabled. lib/streamaggr corrects increase and total results when enable_windows is set and a push carries samples from two windows. lib/httputil sets the Host header to the SRV target for srv+ URLs, lib/netutil uses net.KeepAliveConfig for keepalive parameters, and vmalert copies labels before relabeling and retries notifier config reload after a failed start. vmanomaly documentation was updated for v1.31.0 and UI v1.10.0, and lib/backup adds a flag for S3-compatible storage without tagging directive support.

Week of September 28, 2026

What shipped

Why it matters

V1.153.0 addresses critical issues including data corruption risks, security vulnerabilities, and storage shutdown problems that affect production deployments. The fixes to request handling, authentication errors, and XSS prevention improve reliability and safety across multiple components.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

V1.153.0 is out with fixes for data race in request retries, proper request cancellation during shutdown, and security improvements in vmui and vmauth. See release notes for details on required actions.

Post for LinkedIn

VictoriaMetrics v1.153.0 is now available. This release fixes critical issues including data corruption risks in request retries, improves graceful shutdown with proper request cancellation, and hardens security with XSS prevention and URL parameter validation. Storage improvements and authentication fixes make deployments more reliable.

Week of September 21, 2026

What shipped

Why it matters

This week includes improvements to alerting reliability, authentication capabilities, and query performance. The addition of OIDC SSO support and Prometheus Remote Write v1 protocol support expand integration options. Performance fixes reduce CPU hotspots and disk read overhead under high load.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

V1.153.0-cluster ships with OIDC SSO support in vmauth, Prometheus Remote Write v1 protocol support in vmstorage, improved alert config reloading, reduced query CPU contention, and numerous performance and reliability fixes.

Post for LinkedIn

VictoriaMetrics v1.153.0-cluster is now available. This release adds OpenID Connect single sign-on support to vmauth, enables vmstorage to accept Prometheus Remote Write v1 protocol ingestion, fixes alert configuration reloading in vmalert, reduces CPU contention in query processing, and includes performance optimizations for disk I/O and HTTP request handling.

Week of September 14, 2026

What shipped

Why it matters

This week includes a security release fixing JWT routing authorization bypass and performance improvements to index search through concurrent per-day operations. Cleanup of scrape job configuration and error handling improvements make agent operations more reliable.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

V1.152.0 released with security fix for JWT-based routing in vmauth and concurrent per-day index search for better performance.

Post for LinkedIn

V1.152.0 brings important security fixes and performance improvements. JWT-based routing authorization bypass is patched, concurrent per-day index search accelerates query performance, and scrape job configuration cleanup ensures proper target management.

Weeks with too little public activity are left out rather than filled in. Last updated 2026-11-09.

Is this your repo? Ask us to remove this page.