InfluxDB OSS v2 release notes

v2.7.10

Bug Fixes

  • Update Flux to address performance regression.

Maintenance

  • Update Flux to 0.195.2.

v2.7.9

Bug Fixes

  • Fix UI build error.

Maintenance

  • Update go toolchain to 1.21.12.

v2.7.8

Bug Fixes

  • Prevent an infinite loop in write path.
  • Fix range() filter behavior for timestamps before 1970-01-01 (UNIX time 0).
  • Update to latest influxdb-templates.

Maintenance

  • Update user interface (UI) with various improvements.

v2.7.7

Features

Bug Fixes

  • Ensure TSMBatchKeyIterator and FileStore close all TSMReaders.
  • Return MergeIterator.Close errors.
  • Ensure GROUP BY queries with an offset that crosses a date time boundary (daylight savings time, British summer time, etc.) do not fail.
  • Preserve time zone information in the Task Scheduler.
  • Prevent the retention service from becoming unresponsive.

Maintenance

  • Update Flux to v0.195.1.
  • Update logrus to 1.9.3.
  • Update golang.org/x/net to v0.23.0.
  • Update protocol buffers to v5.26.1.
  • Update go toolchain to 1.21.10.

v2.7.6

Features

Bug Fixes

  • Fix panic index out of range error for invalid series keys.
  • Update broken Flux and performance tests.
  • Do not panic when querying empty tags.
  • Improve shard deletion.
  • Return and respect cursor errors.
  • Update constant time code to make password strength calculations more constant.

Maintenance

  • Upgrade to Go to 1.21.9.

v2.7.5

Features

  • Add authenticating ID and user ID to request logging.
  • Write detailed logs from endpoint detection and response (EDR) failures.

Bug Fixes

  • Only execute init_config on install.
  • Enable Secure when using TLS and enable HttpOnly.
  • Correctly return 4xx errors instead of 5xx errors.
  • Prevent retention service from creating orphaned shard files.

Maintenance

  • Emit build commands during tests.
  • Upgrade Flux to 0.194.5.

v2.7.4

Internal changes only.

v2.7.3

Maintenance

  • Upgrade Flux to 0.194.3.

v2.7.1

Bug Fixes

  • Update the InfluxDB UI to remove non-functional Data Explorer.

v2.7.0

Bug Fixes

  • Update user interface to resolve dashboard crashes and All Access token creation.
  • Show replication queue size and replication TCP errors.
  • Correctly handle when the replication queue cannot be found.
  • Rename stored replication fields to be more clear.
  • Prevent unauthorized writes from Flux to() function.
  • Add error message when attempting to delete by field.
  • Prevent scrapers from failing when the Content-Type header is not set.
  • Removed timeout reached with large databases or slow servers.
  • Changing file permissions on the manifest is no longer necessary.

Maintenance

  • Update Flux to v0.193.0.
  • Update package repository GPG key/
  • Add additional error logging when deleting a shard.
  • Update to Go 1.20.1.
  • Update end-to-end tests to work locally in UI repo.
  • Bump github.com/aws/aws-sdk-go from 1.30.12 to 1.33.0.
  • Bump testcontainers-go to 0.18.0.
  • Bump golang.org/x/net from 0.5.0 to 0.7.0.
  • Bump containerd to 1.6.18.
  • Bump github.com/opencontainers/runc from 1.1.3 to 1.1.5.

v2.6.1

Bug Fixes

  • Update user interface (UI) to fix dashboard page crash issue.
  • Fix All Access token creation issue.

v2.6.0

Features

Bug Fixes

  • Optimize SHOW FIELD KEY CARDINALITY InfluxQL command.
  • Correctly convert replication maximum age setting to seconds.
  • Fix internal metadata migrations when a replication remote already exists.
  • Update /me and /users routes to match documentation.
  • By default, restrict file permissions to 0750 (drwxr-x---).
  • Correctly handle NaN values returned from InfluxDB scrapers.
  • Pin the user interface (UI) to version that addresses dashboard issues.

Maintenance

  • Upgrade to Go 1.18.9.
  • Upgrade Flux to v0.191.0.

v2.5.1

Bug fixes

  • Fix permissions issue in Debian and Red Hat package managers.

v2.5.0

Features

  • Updated user interface (UI).
  • Allow for incremental changes to fields.idx.

Security updates

  • Restriction to view tokens only once, immediately after creation.
  • Set SameSite=strict on session cookie.

Bug fixes

  • Manually scheduled tasks run when expected.
  • Fix error where virtual DBRP mappings were being ignored.
  • Ability to back up all buckets in your InfluxDB instance.
  • Improved speed of deletes when measurement is part of the predicate.
  • Improved error messaging when attempting to create user with an ID that already exists.

Maintenance

v2.4.0

Features

  • Support bucket names in the replications API and allow InfluxDB 1.x DBRP names as bucket names.
  • Add the InfluxQL shell (REPL).
  • Change to operator model so that admin user has instance-level permissions without operator token.

Flux advancement highlights

Bug fixes

  • Log the log level at startup.
  • Fix panic from CreateIterator in Flux queries.
  • Fix error-caching in bufio.Writer.
  • Remove MATCHER_BEHAVIOR environment variable.

Maintenance

v2.3.0

This release includes the following maintenance, features, security updates and bug fixes.

Production ready: replicate data remotely

InfluxDB 2.2 introduced a technical preview of how to replicate data from InfluxDB OSS to InfluxDB Cloud. This feature is now production ready.

Maintenance

Features

  • Add the --instance-id flag to influxd runtime to add the _instance_id tag to remote replications metrics. Helps to identify where the metrics came from, and ensures that two instances don’t overwrite each others’ data.
  • Add signifier to replication user-agent to signify when replication write requests are made. Helps to identify metrics written via replications.

Flux advancement highlights

Performance improvements
  • Vectorize arithmetic operators in map().
  • Vectorize logical operations in map().
  • Enable movingAverage() and cumulativeSum() optimizations by default.
Other highlights
  • Add preview() to experimental package for limiting return rows and tables (as opposed to just rows with limit()).
  • Add date.scale() to let users dynamically scale durations in dates.
  • Add OpenTracing spans to Flux transformations. Lets you monitor Flux scripts more precisely.
  • Add trace option to Flux CLI.
  • Rename addDuration() to add and subDuration() to sub, and moved both of these functions from the experimental package to the date package.
  • Add location support to date.truncate().
  • Add _status tag to PagerDuty record.
  • Refactor operator profile to be in the query statistics.

Security updates

Several security issues were fixed in dependencies and the toolchain used to build InfluxDB, including:

  • An issue in the gopkg.in/yaml.v3 package import that could lead to a DoS in the templates service.
  • An issue in the github.com/buger/jsonparser package import that could potentially lead to a DoS in storage authorization.
  • Cumulative security fixes for Go 1.17.8-1.18.3 are included in this release. These fixes resolve the following InfluxDB issues:
  • An issue with processing large PEM files that could lead to a DoS in the templates service or flux connections using to().
  • An issue in TLSv1.3 and a lack of ticket randomness.
  • A minor issue with filepath.Clean() on Windows.

Bug fixes

  • Fix race condition when remote replicating deadlocks the remote writer that prevented writes to the target.
  • Resolve error when creating v1 authorization (v1 auth) with a nonexistent bucket ID.
  • Add fields to the _tasks bucket to match schema of the same bucket in InfluxDB Cloud. Provides consistency for clients accessing both.
  • Fix rare case where measurement cardinality reported less than zero.
  • Resolve panic on cleaning up failed iterators.

v2.2.0

This release includes the following new features and several bug fixes.

Features

Technical preview: replicate data remotely

Add the option to replicate data from InfluxDB OSS to InfluxDB Cloud.

On rare occasions, remote write failures may cause data in the replication queue to get stuck. To ensure data is not dropped, restart the InfluxDB instance replicating data.

To assess whether this issue is occurring, we recommend periodically doing one of the following:

  • Verify your data is successfully replicated to your target bucket.
  • View your queue size using influx replication list for unexpected growth.

Replicating data remotely lets you do following:

  • Create a durable subscription on a per bucket basis for the purposes of replicating data on-write from an InfluxDB OSS instance to InfluxDB Cloud.
  • Store, analyze, and aggregate data locally while also forwarding newly arriving data to a centralized InfluxDB Cloud account.
  • Configure a maximum buffer size and data age restrictions to restrict the amount of data stored on disk.

Flux updates

  • Update to Flux v0.150.0.
  • Add option to log Flux queries that are cancelled because of server shutdown.

Build maintenance

  • Upgrade protobuf library.

Task metadata

  • Add option to pass type=basic to /api/v2/tasks to return task metadata without the query text.

Troubleshoot with new metrics

To assist with troubleshooting and operational management, expose the following metrics in Prometheus Exposition Format.

TSM compaction metrics
  • storage_compaction
  • storage_compactions_active
  • storage_compactions_duration_seconds_bucket
  • storage_compactions_duration_seconds_count
  • storage_compactions_duration_seconds_sum
  • storage_compactions_queued
  • storage_compactions_total
Failed to add to replication queue metrics
  • PointsFailedToQueue
  • BytesFailedToQueue
Write status metrics
  • pointsWriteRequested
  • pointsWriteOk
  • pointsWriteDropped
  • pointsWriteErr
  • timeout
Disk size metrics per shard
  • total: number of files per shard
  • disk_bytes: data size in bytes for each shard
Cache subsystem metrics
  • inuse_bytes: Current memory consumption of cache
  • diskBytes: Size of most recent snapshot”
  • latest_snapshot: Unix time of most recent snapshot
  • writes_total: Counter of all writes to cache
  • writes_err: Counter of failed writes to cache
  • writes_dropped: Counter of writes (with any dropped points) to cache
WAL subsystem metrics
  • size: Size of WAL in bytes
  • writes: Number of write attempts to the WAL
  • writes_err: Number of failed write attempts to the WAL

See InfluxDB OSS metrics for additional detail.

Display runtime configuration in use by influxd

  • Add api/v2/config endpoint to display the runtime configuration (for example, when you run influxd print-config). This endpoint lets you review runtime configuration while the instance is live.

Recover user credentials

To recover user credentials, use influx recovery user list to retrieve a list of users, and then use influx recovery user update to update the password for a specified user.

Security updates

  • Add the hardening-enabled option to limit flux/pkger HTTP requests. The new hardening-enabled option ensures that InfluxDB first verifies the IP address of the URL is not private. By default, Flux HTTP and template fetching requests are allowed to access localhost and private IP addresses.
  • Disable use of jsonnet with /api/v2/templates/apply. This prevents crafted authenticated requests from exfiltrating files accessible to the user InfluxDB runs as.
  • Add read permissions check for querying data. This prevents authenticated requests using a write-only token from reading data via the InfluxQL /query compatibility API.
  • Add write permissions check for DELETE and DROP MEASUREMENT. This prevents authenticated requests using a read-only token from deleting data via the InfluxQL /query compatibility API. Additionally, several security issues were fixed in dependencies and the toolchain used to build InfluxDB, including:
  • The following cumulative security fixes for Flux v0.161.0 since 0.139.0 are included in this release:
  • Quote db identifiers. This addresses injection vulnerabilities in database connections using to().
  • Make substring check bounds correctly. This prevents authenticated queries from crashing the Flux engine.
  • The cumulative security fixes for Go 1.17.8 since Go 1.17.2 are included in this release. This addresses an issue in the InfluxDB test suite.

Bug fixes

  • Ensure manual task runs can happen concurrently.
  • Extend snapshot copy of backup to filesystems that do not support hard links.
  • Detect misquoted tag values and return an error.
  • Fix potential deadlock in influxd inspect dump-tsi.
  • Successfully handle errors returned by Sketch.Merge.
  • Return X-version and X-Build headers for all requests to /ping endpoint.
  • Add error when meta.db is missing.
  • Sync TSI index file before closing.
  • Fix race condition that could cause restore command to fail.
  • To successfully restore a backup, use copy when a rename spans volumes.
  • Disable use of jsonnet with /api/v2/templates/apply.
  • Ensure that updating a check (/checks/{checkID}) does not require an owner ID.
  • Remove NATS for scraper processing. Note, this fix does not alter scraper functionality–scrapers still work as they did before.
  • nats-port and nats-max-payload-bytes flags have been deprecated.
  • NATS is no longer embedded in InfluxDB. Because InfluxDB no longer requires a port for NATS, port conflict issues are reduced.
  • Resolve the issue that prevented the browser from tracking the cookie expiry correctly, causing the cookie to expire automatically when restarting the browser or changing tabs. Now, the cookie is correctly preserved.
  • Allow unlimited Flux HTTP calls. Previously, HTTP requests failed silently after 100 MB of data transfer.
  • Remove pagination limits on the /telegrafs API. Previously, pagination wasn’t exposed to the API, so API requests were limited to the default 20 pages.

v2.1.1

To address underlying installation package issues, we bumped the 2.1 release version to 2.1.1.

This release includes several new features and bug fixes.

influx CLI moved to separate repository

The influx CLI has been moved to its own GitHub repository.

influxdb release artifacts affected

Release artifacts produced by influxdb are impacted as follows:

  • Release archives (.tar.gz and .zip) no longer contain the influx binary.
  • The influxdb2 package (.deb and .rpm) no longer contains the influx binary. Instead, the package declares a recommended dependency on the new influxdb2-cli package.
  • The quay.io/influxdb/influxdb image no longer contains the influx binary. We recommend migrating to the influxdb image hosted on DockerHub.

Find supported influx CLI commands

With this change, versions of the influx CLI and InfluxDB server (influxd) are not guaranteed to exactly match. To check the version of the influxd server, see influxd version in influx CLI or use the /health endpoint of your InfluxDB server.

The influx CLI documentation has been updated to reflect which influx CLI commands work with which versions of InfluxDB.

Features

This release includes the following new features:

Notebooks, annotations, and visualization updates

  • Add support for notebooks and annotations.
  • Add support for static legends to line graphs and band plots.
  • Enable new dashboard auto-refresh.
  • Simplify display of data for table visualizations.

API

CLI

influxd configuration

Added several new configuration options to influxd:

  • Add influxd recovery command to let you create a recovery Operator token.
  • Add --sqlite-path flag for specifying a user-defined path to the SQLite database file.
  • Add --storage-wal-max-concurrent-writes flag to enable tuning memory pressure under heavy write load.
  • Add --storage-wal-max-write-delay flag to prevent deadlocks when the WAL is overloaded with concurrent writes.
  • Add --storage-write-timeout flag to set write request timeouts.
  • Add --storage-no-validate-field-size flag to disable enforcement of max field size.
  • Update --store flag to work with string values disk or memory. Memory continues to store metadata in-memory for testing; disk persists metadata to disk via bolt and SQLite.

For more information, see InfluxDB configuration options.

influxd inspect

Ported the following influxd inspect commands from InfluxDB 1.x:

influxd downgrade

Added the influxd downgrade command to migrate InfluxDB key-value metadata schemas to earlier 2.x versions when necessary.

Flux

  • Update to Flux v0.139.0.
  • Enable writing to remote hosts using the Flux to() and experimental.to() functions.
  • Flux now supports locations that dynamically modify time offsets based on your specified timezone. You can also specify fixed time offsets relative to UTC.
  • Perform bitwise operations on integers and unsigned integers.
  • Query and write to Vertica SQL databases.
  • Add the hex package for working with hexadecimal string values.

InfluxQL

  • SHOW MEASUREMENTS ON now supports database and retention policy wildcards. For example, SHOW MEASUREMENTS ON *.* to show all databases and SHOW MEASUREMENTS ON <db>.* to show all retention policies.
  • Add hyper log operators merge_hll, sum_hll, and count_hll in InfluxQL to optimize series iteration. (count_hll optimizes queries that can be answered without inspecting TSM data.)

Telegraf

  • Add the following new Telegraf plugins to the Load Data page:
  • Alibaba (Aliyun) CloudMonitor Service Statistics (aliyuncms)
  • AMD ROCm System Management Interface (SMI) (amd_rocm_smi)
  • Counter-Strike: Global Offensive (CS:GO) (csgo)
  • Directory Monitoring (directory_monitor)
  • Intel Data Plane Development Kit (DPDK) (dpdk)
  • Elasticsearch Query (elasticsearch_query)
  • Internet Speed Monitor (internet_speed)
  • KNX (knx_listener)
  • Mdstat (mdstat)
  • Netstat (netstat)
  • NFS Client (nfsclient)
  • OpenTelemetry (opentelemetry)
  • RavenDB (ravendb)
  • SQL (sql)
  • Trig (trig)
  • Twemproxy (twemproxy)

For more information about each plugin, see Telegraf plugins, and search for the plugin name.

Tokens

  • Add support for standard Bearer token syntax. Now you can specify token credentials as: Authorization: Bearer xxxxxxxx.
  • If restoring a backup overwrites the Operator token, the new token value is returned.

Bug fixes

  • Log API errors to server logs and tell clients to check the server logs for the error message.
  • Fix pagination for GET /buckets API when displaying results. Previously, pagination was broken if a request included both an org filter AND the after request parameter. Also corrects descending parameter to sort when an org filter is used and saved.
  • Sync series segment to disk after writing.
  • Do not allow shard creation to create overlapping shards.
  • Don’t drop shard group durations when upgrading InfluxDB.

v2.0.9

This release includes several new features and bug fixes.

Features

New features include:

API updates

  • Add a new route /api/v2/resources that returns a list of known resources to the platform, including the following resource types. Makes it easier to update All Access tokens with current resources:

  • AuthorizationsResourceType

  • BucketsResourceType
  • ChecksResourceType
  • DashboardsResourceType
  • DBRPResourceType
  • DocumentsResourceType
  • LabelsResourceType
  • NotificationEndpointResourceType
  • NotificationRuleResourceType
  • OrgsResourceType
  • ScraperResourceType
  • SecretsResourceType
  • SourcesResourceType
  • TasksResourceType
  • TelegrafsResourceType
  • UsersResourceType
  • VariablesResourceType
  • ViewsResourceType

Flux updates

  • Update to Flux v0.130.0.
  • Add support for influxdb.cardinality() function.
  • Operational improvements:
  • Add logging to Flux end-to-end tests (TestFluxEndToEnd) to help diagnose test failures.
  • Add --flux-log-enabled option to influxd to show detailed logs for Flux queries.

Performance enhancements

  • Optimize series iteration for queries that can be answered without inspecting TSM data.
  • Optimize queries with predicates that contain multiple measurements.

Bug fixes

This release includes the following bug fixes and updates:

API fix

  • Correctly filter requests to /api/v2/authorizations by org and user parameters.

Dependency update

  • Include curl as a dependency in influxdb2 packages.

Errors updates

  • Add message to set the Flux content-type when a query fails to parse as JSON.
  • Discard excessive errors over DefaultMaxSavedErrors (100) to prevent out-of-memory crashes.
  • Upgrade golang.org/x/sys to avoid panics on macs.

Limit update

  • Implement hard limit on field size (MaxFieldValueLength = 1048576) while parsing line protocol.

Miscellaneous operational fixes

  • Resolve the compaction queue stats flutter.
  • Ensure the TSI index compacts log files that meet one of the following criteria:
  • Log file hasn’t been updated (no new series have been added to the shard) for 4 (or more) hours (to change this duration, specify a new storage-compact-full-write-cold-duration)
  • Log file is one (or more) megabytes (to update this size, specify a new storage-max-index-log-file-size)
  • Repair bad port dropping return value names.
  • Use consistent path separator in permission string representation.
  • (Windows only) Copy snapshot files being backed up.

Task updates

  • Updating an inactive task no longer schedules it.
  • Preserve comments in Flux queries when saving task definitions.

Version maintenance

  • Fix X-Influxdb-Build and X-Influxdb-Version response header at /ping.
  • Upgrade influxql to latest version and fix predicate handling for SHOW TAG VALUES meta queries.

v2.0.8

Upcoming changes to influx CLI packaging

Beginning in InfluxDB 2.1, the influx CLI will no longer be packaged with the release. Future versions of influx CLI will be released from the influx-cli repository.

To adopt the new, separate influx CLI early, download the latest release from GitHub or from the InfluxData Downloads portal.

Go version

  • Upgrade to Go 1.16. Requires macOS Sierra 10.12 or later to run.

Features

  • Add --ui-disabled option to influxd to support running with the UI disabled.
  • Telemetry improvements: Do not record telemetry data for non-existent paths; replace invalid static asset paths with a slug.
  • Upgrade to Flux v0.124.0.
  • Upgrade to UI v2.0.8.
  • Upgrade flux-lsp-browser to v0.5.53.

Bug fixes

  • Rename ARM RPM packages with yum-compatible names.
  • Upgrade to latest version of influxdata/cron so that tasks can be created with interval of every: 1w.
  • Avoid rewriting fields.idx unnecessarily.
  • Do not close connection twice in DigestWithOptions.
  • Remove incorrect optimization for group-by.
  • Return an error instead of panicking when InfluxQL statement rewrites fail.
  • Migrate restored KV snapshots to latest schema before using them.
  • Specify which fields are missing when rejecting an incomplete onboarding request.
  • Ensure systemd unit blocks on startup until HTTP endpoint is ready.
  • Fix display and parsing of influxd upgrade CLI prompts in PowerShell.
  • Removed unused chronograf-migator package and chronograf API service, and updated various “chronograf” references.
  • Fix display and parsing of interactive influx CLI prompts in PowerShell.
  • Upgrade to golang-jwt 3.2.1.
  • Prevent silently dropped writes when there are overlapping shards.
  • Invalid requests to /api/v2 subroutes now return 404 instead of a list of links.
  • Flux meta queries for _field take fast path if _measurement is the only predicate.
  • Copy names from mmap mapped memory before closing iterator.

v2.0.7

Features

  • Optimize table.fill() execution within Flux aggregate windows.
  • Upgrade Flux to v0.117.0.
  • Upgrade UI to v2.0.7.
  • Upgrade flux-lsp-browser to v0.5.47.

Bug Fixes

  • Fix query range calculation (off by one) over partially compacted data.
  • Deprecate the unsupported PostSetupUser API.
  • Add limits to the /api/v2/delete endpoint for start and stop times with error messages.
  • Add logging to NATS streaming server to help debug startup failures.
  • Accept --input instead of a positional argument in influx restore.
  • Print error instead of panicking when influx restore fails to find backup manifests.
  • Set last-modified time of empty shard directory to the directory’s last-modified time, instead of the Unix epoch.
  • Remove deadlock in influx org members list when an organization has greater than 10 members.
  • Replace telemetry file name with slug for ttf, woff, and eot files.
  • Enable use of absolute path for --upgrade-log when running influxd upgrade on Windows.
  • Make InfluxQL meta queries respect query timeouts.

v2.0.6 General Availability

Bug Fixes

  • Ensure query configuration written by influxd upgrade is valid.
  • Set query-concurrency and query-queue-size configuration option defaults to 0 to avoid validation failures when upgrading users.
  • Correctly validate when query-concurrency is 0 and query-queue-size is greater than 0.

v2.0.5 General Availability

InfluxDB v2.0.5 introduced a defect that prevents users from successfully upgrading from InfluxDB 1.x to 2.0 using the influxd upgrade command or Docker. To automatically upgrade from 1.x to 2.0 with the influxd upgrade command or with Docker, use InfluxDB v2.0.6.

Windows Support

This release includes our initial Windows preview build.

Breaking Changes

/debug/vars removed

Prior to this release, the influxd server would expose profiling information over the /debug/vars endpoint. This endpoint was unauthenticated and not used by InfluxDB systems to report diagnostics. For security and clarity, the endpoint has been removed. Use the /metrics endpoint to collect system statistics.

influx transpile removed

The transpile command has been removed. Send InfluxQL requests directly to the server via the /api/v2/query or /query HTTP endpoints.

Default query concurrency changed

The default setting for the max number of concurrent Flux queries has been changed from 10 to unlimited (0). To limit query concurrency and queue size:

  1. Set the query-concurrency config parameter to > 0 when running influxd to re-limit the maximum running query count,
  2. Set the query-queue-size config parameter to > 0 to set the max number of queries that can be queued before the server starts rejecting requests.

Prefix for query-controller metrics changed

The prefix used for Prometheus metrics from the query controller has changed from query_control_ to qc_.

Features

Bug Fixes

  • Prevent “do not have an execution context” error when parsing Flux options in tasks.
  • Fix swagger to match implementation of DBRPs type.
  • Fix use-after-free bug in series ID iterator.
  • Fix TSM and WAL segment size check to check against the local SegmentSize.
  • Fix TSM and WAL segment size computing to correctly calculate totalOldDiskSize.
  • Update references to the documentation site to use current URLs.
  • Fix data race in then TSM engine when inspecting tombstone statistics.
  • Fix data race in then TSM cache.
  • Deprecate misleading retentionPeriodHrs key in the onboarding API.
  • Fix Single Stat graphs with thresholds crashing on negative values.
  • Fix InfluxDB port in Flux function UI examples.
  • Remove unauthenticated, unsupported /debug/vars HTTP endpoint.
  • Respect 24 hour clock formats in the InfluxDB UI and add more format choices.
  • Prevent “do not have an execution context” error when parsing Flux options in tasks.
  • Prevent time field names from being formatted in the Table visualization.
  • Log error details when influxd upgrade fails to migrate databases.
  • Fix the cipher suite used when TLS strict ciphers are enabled in influxd.
  • Fix parse error in UI for tag filters containing regular expression meta characters.
  • Prevent concurrent access panic when gathering bolt metrics.
  • Fix race condition in Flux controller shutdown.
  • Reduce lock contention when adding new fields and measurements.
  • Escape dots in community templates hostname regular expression.

v2.0.4 General Availability

Docker

ARM64

This release extends the Docker builds hosted in quay.io to support the Linux/ARM64 platform.

2.x nightly images

Prior to this release, competing nightly builds caused the nightly Docker tag to contain outdated binaries. This conflict is fixed, and the image tagged with nightly now contains 2.x binaries built from the HEAD of the master branch.

Breaking Changes

inmem index option removed

This release fully removes the inmem indexing option, along with the associated config options:

  • max-series-per-database
  • max-values-per-tag

The startup process automatically generates replacement tsi1 indexes for shards that need it.

Features

influxd updates

New Telegraf plugins in UI

  • Update Telegraf plugins list in UI to include Beat, Intel PowerStats, and Rienmann.

Performance improvements

  • Promote schema and fill query optimizations to default behavior.

Flux updates

Bug Fixes

  • Standardize binary naming conventions.
  • Fix configuration loading issue.
  • Add Flux dictionary expressions to Swagger documentation.
  • Ensure influxdb service sees default environment variables when running under init.d.
  • Remove upgrade notice from new installs.
  • Ensure config.toml is initialized on new installs.
  • Include upgrade helper script (influxdb2-upgrade.sh) in GoReleaser manifest.
  • Prevent influx stack update from overwriting stack name and description.
  • Fix timeout setup for influxd graceful shutdown.
  • Require user to set password during initial user onboarding.
  • Error message improvements:
  • Remove duplication from task error messages.
  • Improve error message shown when influx CLI can’t find an org by name.
  • Improve error message when opening BoltDB with unsupported file system options.
  • Improve messages in DBRP API validation errors.
  • influxd upgrade improvements:
  • Add confirmation step with file sizes before copying data files.
  • Prevent panic in influxd upgrade when v1 users exist but v1 config is missing.
  • Fix logging initialization for storage engine.
  • Don’t return 500 codes for partial write failures.
  • Don’t leak .tmp files while backing up shards.
  • Allow backups to complete while a snapshot is in progress.
  • Fix silent failure to register CLI arguments as required.
  • Fix loading when INFLUXD_CONFIG_PATH points to a .yml file.
  • Prevent extra output row from GROUP BY crossing DST boundary.
  • Update Flux functions list in UI to reflect that v1 package was renamed to schema.
  • Set correct Content-Type on v1 query responses.
  • Respect the --skip-verify flag when running influx query.
  • Remove blank lines from payloads sent by influx write.
  • Fix infinite loop in Flux parser caused by invalid array expressions.
  • Support creating users without initial passwords in influx user create.
  • Fix incorrect errors when passing --bucket-id to influx write.

v2.0.3 General Availability

Breaking Changes

influxd upgrade

Previously, influxd upgrade would attempt to write upgraded config.toml files into the same directory as the source influxdb.conf file. If this failed, a warning would be logged and config.toml would write into the home directory of the user who launched the upgrade.

This release breaks this behavior in two ways:

  • By default, config.toml writes into the same directory as the Bolt DB and engi

（本文超出站点单文档体积上限，已截断）