Shipped · Databases

What influxdata/influxdb shipped

The public repository of Influxdata · github.com/influxdata/influxdb

Written by FoxPlug from the repository's own releases; not affiliated with Influxdata. 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/influxdata/influxdb; Influxdata 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 influxdb, free

Week of November 2, 2026

What shipped

Why it matters

This week tightens the storage engine's handling of concurrent series changes, which affects anyone running queries, deletes or DROP MEASUREMENT against the inmem and TSI indexes. It also cuts allocation load when tombstoning many series. Two InfluxDB 3 releases bring a catalog repair command to 3.12.2 and Python and memory-limit query fixes to 3.11.8, and both 1.11 and 1.12 move to Go 1.26.10.

Changelog entry

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

Post for X

This week: index race fixes for inmem and TSI, fewer allocations when tombstoning series, v3.12.2 and v3.11.8 out, and Go 1.26.10 on 1.11 and 1.12.

Post for LinkedIn

This week's changes to InfluxDB: the inmem index no longer returns duplicate or missing series when the series list changes during reads, and DROP MEASUREMENT no longer crashes the server under the TSI index. SeriesIDSet iterators are now closed only when the data is no longer in use, and the gzip writer is re-used across tombstone batches, cutting total allocations for a DROP SERIES measurement from 37.87 GiB to 11.03 GiB. In InfluxDB 3, v3.12.2 adds the influxdb3 debug catalog repair command and v3.11.8 updates Python to 3.13.18 and fixes memory-limit queries returning 500. InfluxDB OSS 1.11 and 1.12 now build with Go 1.26.10.

Week of October 26, 2026

What shipped

Why it matters

This week brings releases on the 1.x and 3.x lines alongside a tsdb fix that ensures measurement deletions are recorded in fields.idxl, which matters for anyone relying on DROP MEASUREMENT, DROP SERIES or DELETE. The Go version updates and dependency syncs keep the branches current, and the new benchmark gives a baseline for the HLL.Clone() investigation.

Changelog entry

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

Post for X

V1.13.2, v3.11.8 and v3.12.2 are out. Measurement deletions from DROP MEASUREMENT, DROP SERIES and DELETE now write to fields.idxl instead of being dropped. Go updated to 1.26.10, dependency syncs for 3.11 and 3.12, and a new store.Statistics benchmark.

Post for LinkedIn

This week we published v1.13.2, v3.11.8 and v3.12.2. A tsdb fix addresses measurement deletions: DROP MEASUREMENT, DROP SERIES and DELETE that remove the last series of a measurement now save DeleteMeasurement field changes with Field == nil, where previously every deletion was dropped because marshalFieldChanges only appended changes with a non-nil Field. We also updated Go to 1.26.10, synced influxdb_pro changes into 3.11 and 3.12, including a pulldown-cmark update to 0.13.5 and a Python update to 3.13.18-20261026 for security fixes, and added a store.Statistics benchmark for the HLL.Clone() investigation.

Week of October 19, 2026

What shipped

Why it matters

This week fixes a real correctness bug for Python processing engine plugins: ON_WRITE table_batches rows could be misbound when a table flush contained multiple RecordBatches with the same width but different column order. Users writing plugins that read WAL data should see columns land under the right names instead of by position. The remaining changes bring the tooling build in line and clean up an unimplemented interface member in influx_tools.

Changelog entry

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

Post for X

This week: Python API WAL RecordBatch columns are now bound by name for ON_WRITE table_batches, fixing misbinding when a table flush has multiple RecordBatches of the same width but different column order. Also: influx_tools interface fix and Go updated to 1.26.10.

Post for LinkedIn

This week in the repository: the Python API now binds WAL RecordBatch columns by name when building ON_WRITE table_batches rows, fixing a case where a table flush with multiple RecordBatches of the same width but different column order could bind columns by position incorrectly. Processing engine plugins that read WAL data benefit from this. Also shipped: a fix for an unimplemented interface member in influx_tools, plus an update to Go 1.26.10, both carried onto the release branch.

Week of October 12, 2026

What shipped

Why it matters

This week brought a new InfluxDB 3.12 release with faster bulk import, distributed compaction in beta, per-database schema enforcement and generally available authentication and role-based access. Two patch releases followed with faster startup after large file removals and a fix for catalogs that failed to load after an upgrade. Security and stability fixes also arrived through rustls updates and serialized Parquet WAL flushes.

Changelog entry

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

Post for X

InfluxDB 3.12 is out: faster bulk import, distributed compaction in beta, per-database schema enforcement, and GA user authentication and role-based access. 3.12.1 and 3.11.7 followed with faster startup after large file removals and a catalog load fix. Build now uses the flux_no

Post for LinkedIn

This week InfluxDB 3.12 shipped with faster bulk import, distributed compaction in beta, per-database schema enforcement, and generally available user authentication and role-based access. Two patch releases followed: 3.12.1 makes startup faster after large file removals, and 3.11.7 fixes a catalog that failed to load after upgrading to 3.11.6 when a table has more than one field family with the same automatic name. Smaller changes landed too: the build now uses the flux_nosqlite flag, the Go dependency moved to 1.26.9, and 1.11 moved to the same Go version. Syncs from the pro repository brought a rustls update for security advisories and made Parquet WAL flushes run one at a time.

Week of September 28, 2026

What shipped

Why it matters

V3.12.0 ships major improvements to health and readiness checks: they now report structured data, support authentication, survive startup failures longer, and name failing subsystems. These changes make it easier for operators to monitor influxd startup and health status.

Changelog entry

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

Post for X

V3.12.0 released. /health and /ready endpoints now report structured JSON fields, support authentication, and name failing subsystems at startup. Exit codes now match sysexits.h on known failures.

Post for LinkedIn

V3.12.0 is out. We've improved health and readiness monitoring: /health and /ready now report numbers as structured JSON fields instead of free-text, support operator-only authentication, name the failing subsystem on startup errors, and hold open after failed startup. Exit codes now follow sysexits.h standards for different failure types. Better observability for operators running influxd.

Week of September 21, 2026

What shipped

Why it matters

This week's updates focus on operational visibility and stability. Users gain new per-user metrics for writes and queries to monitor usage patterns, adaptive caching to reduce manual tuning, and query limits to prevent resource exhaustion. Several fixes address race conditions and data integrity issues in core compaction and shard loading operations.

Changelog entry

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

Post for X

V3.12.0-rc.2 ships with per-user write and query metrics, adaptive caching bounds, compaction planning stats, and critical fixes for shard loading races and TSM level regressions.

Post for LinkedIn

V3.12.0-rc.2 is out with new operational features: per-user write and query byte statistics for usage monitoring, adaptive TagValueSeriesIDCache sizing to reduce configuration overhead, and detailed compaction planning statistics. This release also fixes critical issues including race conditions in shard loading and level regressions during mixed-level TSM compaction, improving both observability and reliability.

More from influxdata/influxdb on GitHub

Merged pull requests

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.