Grafana Release Notes

BREAKING CHANGES

• prometheus.operator.servicemonitors: prometheus.operator.servicemonitors now rejects ServiceMonitor endpoints that reference local files through bearerTokenFile, tlsConfig.caFile, tlsConfig.certFile, or tlsConfig.keyFile by default. Set allow_arbitrary_file_access = true to preserve the previous behavior for trusted ServiceMonitor authors.
• The memory_limiter processor's internal telemetry metrics were renamed upstream in OpenTelemetry Collector v0.155.0 to carry a memory_limiter prefix
• Remove the prometheus.write.queue component (#6814)

Features 🌟

• Add ability to exempt lines matching a configured regex pattern from loki.process stage.luhn (#6608) (2e0748d)
• Add opamp supervisor (#6815) (e891e30)
• Add start_timestamp_zero_ingestion to Prometheus metric components (#6610) (1a9ed95)
• beyla.ebpf: Run Beyla as a subprocess (#6695) (61ec3d2)
• Binary Alloy Homebrew Formula (#6723) (024d224)
• Bump otel to v0.158.0 and prometheus to v3.13.2 (#6842) (aef2e8b)
• Check imports of inline Alloy config in OTel mode (#6677) (fda0449)
• database_observability.mysql: Add support for clustering mode (#6775) (0cfa960)
• database_observability.postgres: Add user, pid, and sqlstate fields to op=error_message (#6770) (73e29dc)
• database_observability.postgres: Emit op=query_association_v2 carrying query_fingerprint (#6297) (75239d0)
• database_observability.sql_server: Add query_details collector (#6788) (cec31b2)
• database_observability.sql_server: Add query_metrics collector (#6743) (35babfd)
• database_observability.sql_server: Support csp metadata for SQL Server (#6689) (24c7207)
• discovery.azure: Add "sdk_auth" and "workload_identity" auth options (#6671) (9fa8b51)
• Expose experimental OTel supervisor (#6903) (936c177)
• Expose new kafka receiver/exporter knobs after upgrade to v0.158.0 (#6868) (88601cb)
• loki.process: Add stage.split_json to split top-level JSON arrays (#6823) (7bdd3dd)
• loki.secretfilter: Promote loki.secretfilter to public preview (#6724) (4053610)
• otelcol.receiver.prometheus: Return stable refs (#6822) (4633351)
• otelcol: Add otelcol.processor.redaction (#6647) (ea35e8a)
• Part 2 - Expose knobs for OTel v0.158.0 upgrade (#6876) (397acb1)
• Part 3 - Expose knobs for OTel v0.158.0 upgrade - add exposed_headers to otel http config (#6889) (c42f0c4)
• Part 4 - Expose knobs for Prometheus 3.13.2 & OTel v0.158.0 upgrade (#6890) (8a877bf)
• prometheus.exporter.mssql: Add a new max_connection_lifetime config attribute (#6700) (5286dd6)
• prometheus.operator: Add missing native histogram scrape for parity with scrape (#6406) (368f772)
• prometheus.operator: Expose bodySizeLimit and enableCompression (#6657) (8fa09a1)
• pyroscope.java: Add tlab profiling config option for TLAB-based allocation events (#6754) (b955613)
• pyroscope.java: Update async-profiler binaries to v4.5 (#6739) (9b8aa87)
• pyroscope.write: Add retry_on_http_429 endpoint option (#6763) (bcd909d)
• pyroscope: Support span context propagation from OBI (including Beyla component) (#6298) (24d66d0)

Bug Fixes 🐛

• Add nop receiver to OTel Engine (#6751) (8b7703c)
• beyla.ebpf: Remove wrong native_histograms config attribute (#6086) (99074ba)
• beyla: Unix socket name could be too long (#6905) (bb7d902)
• Bump go version to v1.26.6 (#6897) (27b94ac)
• Clashing cross compilation env variables when building/generating alloy (#6793) (0c7af36)
• database_observability.postgres: Restore cgo SQL fingerprinting on Windows (#6746) (bc33331)
• database_observability.postgres: Roll back op=query_association_v2, fold fingerprint into op=query_association (#6745) (40e8f1e)
• database_observability.postgres: Use of defer in explain_plans collector (#6819) (9935420)
• dependencies: Upgrade golang.org/x/text to v0.39.0 (#6744) (105ecf6)
• discovery.relabel: Use the underlying string of secret target values (#6605) (171aea1)
• fmt: Do not write a file if no formatter changes are needed (#6871) (a7696d5)
• Harden the configmap reloader container (#6780) (e49e090)
• loki.process: Return errors from metric registration instead of panicking (#6078) (a388da5)
• loki.source.docker: Respect label changes (#6818) (d0525a9)
• loki.source.docker: Restart tailers when config changes (#6826) (eb7f3ef)
• loki.source.docker: Stop tailer when container no longer exists (#6309) (44f0377)
• loki.source.file, loki.source.journal: Fail loudly on unreadable legacy positions file (#6750) (2425b8a)
• loki.source.windowsevent: Use octal notation for file permissions (#6077) (789140d)
• mixin: Fix overlapping panels and inconsistent selector application for loki and OTel Engine (#6947) (66ebea8)
• otelcol.exporter.loadbalancing: Change default Kubernetes resolver timeout to 1m (#6702) (d425ad9)
• prometheus.fanout: Correct series-ref handling on the unmapped fanout path (#6821) (2edaaab)
• prometheus.operator.servicemonitors: Sync ServiceMonitor permission semantics with upstream (2eeaa3e)
• prometheus.receive_http: Eliminate free-port race in tests (#6827) (37b652c)
• prometheus: Reduce metadata pipeline allocation overhead (#6698) (1fef867)
• security/HIGH/internal/web/ui: Update npm dependency react-router to v7.18.2 [SECURITY] (#6848) (d78bd74)
• security/UNKNOWN/tools: Update module golang.org/x/mod to v0.40.0 [SECURITY] (#6888) (231c259)
• security/UNKNOWN/: Update module golang.org/x/mod to v0.40.0 [SECURITY] (#6886) (6e397ea)
• security/UNKNOWN/: Update module google.golang.org/grpc to v1.82.1 [SECURITY] (#6735) (bffcf59)
• ui: Move legends to bottom (#6851) (546c65f)
• Upgrade go-re2 to v1.12.0 to allow MemoryDenyWriteExecute=true with loki.secretfilter (#6737) (f34140a)
• Use default upstream config for vcenter receiver (#6918) (107a9eb)

Chores

• Remove the prometheus.write.queue component (#6814) (b168d0e)

Breaking changes due to bugfixes in Prometheus exporters

The prometheus.exporter.blackbox, prometheus.exporter.snmp, and prometheus.exporter.statsd components now use the component ID instead of the hostname as their instance label in their exported metrics. This is a consequence of a bug fix that could lead to missing data when using the exporter with clustering.

If you would like to retain the previous behaviour, you can use discovery.relabel with action = "replace" rule to set the instance label to sys.env("HOSTNAME").

Breaking changes due to major version upgrade of Prometheus

Prometheus dependency had a major version upgrade from v2.55.1 to v3.4.2.

• The . pattern in regular expressions in PromQL matches newline characters now. With this change a regular expressions like .* matches strings that include \n. This applies to matchers in queries and relabel configs in Prometheus and Loki components.

• The enable_http2 in prometheus.remote_write component’s endpoints has been changed to false by default. Previously, in Prometheus v2 the remote write http client would default to use http2. In order to parallelize multiple remote write queues across multiple sockets its preferable to not default to http2. If you prefer to use http2 for remote write you must now set enable_http2 to true in your prometheus.remote_write endpoints configuration section.

• The experimental CLI flag --feature.prometheus.metric-validation-scheme has been deprecated and has no effect. You can configure the metric validation scheme individually for each prometheus.scrape component.

• Log message format has changed for some of the prometheus.* components as part of the upgrade to Prometheus v3.

• The values of the le label of classic histograms and the quantile label of summaries are now normalized upon ingestion. In previous Alloy versions, that used Prometheus v2, the value of these labels depended on the scrape protocol (protobuf vs text format) in some situations. This led to label values changing based on the scrape protocol. E.g. a metric exposed as my_classic_hist{le="1"} would be ingested as my_classic_hist{le="1"} via the text format, but as my_classic_hist{le="1.0"} via protobuf. This changed the identity of the metric and caused problems when querying the metric. In current Alloy release, which uses Prometheus v3, these label values will always be normalized to a float like representation. I.e. the above example will always result in my_classic_hist{le="1.0"} being ingested into Prometheus, no matter via which protocol. The effect of this change is that alerts, recording rules and dashboards that directly reference label values as whole numbers such as le="1" will stop working.

The recommended way to deal with this change is to fix references to integer le and quantile label values, but otherwise do nothing and accept that some queries that span the transition time will produce inaccurate or unexpected results.

See the upstream Prometheus v3 migration guide for more details.

Breaking changes in prometheus.scrape

scrape_native_histograms attribute for prometheus.scrape is now set to false, whereas in previous versions of Alloy it would default to true. This means that it is no longer enough to just configure scrape_protocols to start with PrometheusProto to scrape native histograms - scrape_native_histograms has to be enabled. If scrape_native_histograms is enabled, scrape_protocols will automatically be configured correctly for you to include PrometheusProto. If you configure it explicitly, Alloy will validate that PrometheusProto is in the scrape_protocols list.

In previous versions of Alloy configuring scrape_protocols to start with PrometheusProto was enough to start scraping native histograms because scrape_native_histogram defaulted to true:

Alloy prometheus.scrape "scrape" { scrape_protocols = ["PrometheusProto"] }

Now it has to be enabled and scrape_protocols can be omitted:

Alloy prometheus.scrape "scrape" { scrape_native_histograms = true }

Breaking changes in prometheus.exporter.windows

As the windows_exporter continues to be refactored upstream, there are various breaking changes in metrics.

• windows_process_start_time -> windows_process_start_time_seconds_timestamp in the process collector.
• windows_time_clock_frequency_adjustment_ppb_total -> windows_time_clock_frequency_adjustment_ppb in the time collector.
• windows_net_nic_info -> windows_net_nic_address_info in net collector.
• windows_system_boot_time_timestamp_seconds -> windows_system_boot_time_timestamp in system collector.
• windows_os_physical_memory_free_bytes -> windows_memory_physical_free_bytes from os collector -> memory collector.
• windows_os_process_memory_limit_bytes -> windows_memory_process_memory_limit_bytes from os collector -> memory collector.
• windows_os_processes -> windows_system_processes from os collector -> system collector.
• windows_os_processes_limit -> windows_system_process_limit from os collector -> system collector.
• windows_os_time -> windows_time_current_timestamp_seconds from os collector -> time collector.
• windows_os_timezone -> windows_time_timezone from os collector -> time collector.
• windows_os_users from os collector can be reconstructed by aggregating windows_terminal_services_session_info{state="active"} in terminal_services collector.
• windows_os_virtual_memory_bytes -> windows_memory_commit_limit from os collector -> memory collector.
• windows_os_virtual_memory_free_bytes from os collector can be reconstructed by subtracting windows_memory_committed_bytes from windows_memory_commit_limit in memory collector.
• windows_os_visible_memory_bytes -> windows_memory_physical_total_bytes from os collector -> memory collector.

Deprecated collectors have been removed. Configuration of these collectors will be allowed in Alloy for at least one minor version, but will have no effect.

• logon collector removed, use the terminal_services collector instead.
• cs collector removed, use the os, memory, or cpu collectors instead.

Refer to the release notes for the windows_exporter version v0.31.0 for the breaking change details.

Breaking change: The prometheus.exporter.oracledb component now embeds a different exporter

Breaking change: The prometheus.exporter.oracledb component now embeds a different exporter

The prometheus.exporter.oracledb component now embeds the oracledb_exporter from oracle instead of the deprecated oracledb_exporter from iamseth.

As a result of this change, the following metrics are no longer available by default:

• oracledb_sessions_activity
• oracledb_tablespace_free_bytes

The previously undocumented argument custom_metrics is now expecting a list of paths to custom metrics files.

Breaking change: The enable_context_propagation argument in beyla.ebpf has been replaced with the context_propagation argument.

Breaking change: The enable_context_propagation argument in beyla.ebpf has been replaced with the context_propagation argument.

Set enable_context_propagation to all to get the same behaviour as enable_context_propagation being set to true.

Breaking change: In prometheus.exporter.windows, the service and msmq collectors no longer work with WMI

Breaking change: In prometheus.exporter.windows, the service and msmq collectors no longer work with WMI

The msmq block has been removed. The enable_v2_collector, where_clause, and use_api attributes in the service block are also removed.

Prior to Alloy v1.9.0, the service collector exists in 2 different versions. Version 1 used WMI (Windows Management Instrumentation) to query all services and was able to provide additional information. Version 2 is a more efficient solution by directly connecting to the service manager, but is not able to provide additional information like run_as or start configuration.

In Alloy v1.9.0 the Version 1 collector was removed, hence why some arguments and blocks were removed. In Alloy v1.9.2 those arguments and blocks were re-introduced as a no-op in order to make migrations easier for customers.

Due to this change, the metrics produced by service collector are different in v1.9.0 and above. The msmq collector metrics are unchanged.

Example V2 service metrics:

windows_service_state{display_name="Declared Configuration(DC) service",name="dcsvc",status="continue pending"} 0
windows_service_state{display_name="Declared Configuration(DC) service",name="dcsvc",status="pause pending"} 0
windows_service_state{display_name="Declared Configuration(DC) service",name="dcsvc",status="paused"} 0
windows_service_state{display_name="Declared Configuration(DC) service",name="dcsvc",status="running"} 0
windows_service_state{display_name="Declared Configuration(DC) service",name="dcsvc",status="start pending"} 0
windows_service_state{display_name="Declared Configuration(DC) service",name="dcsvc",status="stop pending"} 0
windows_service_state{display_name="Declared Configuration(DC) service",name="dcsvc",status="stopped"} 1

For more information on V1 and V2 service metrics, see the upstream exporter documentation for version 0.27.3 of the Windows Exporter, which is the version used in Alloy v1.8.3. Alloy v1.9.2 uses version 0.30.7 of the Windows Exporter.

Breaking change: The topics argument in the component loki.source.kafka does not use regex by default anymore

Breaking change: The topics argument in the component loki.source.kafka does not use regex by default anymore

A bug in loki.source.kafka caused the component to treat all topics as regular expressions. For example, setting the topic value to “telemetry” would match any topic containing the substring “telemetry”. With the fix introduced in this version, topic values are now treated as exact matches by default. Regular expression matching is still supported by prefixing a topic with “^”, allowing it to match multiple topics.

Breaking change: Change decision precedence in otelcol.processor.tail_sampling when using and_sub_policy and invert_match

Breaking change: Change decision precedence in otelcol.processor.tail_sampling when using and_sub_policy and invert_match

Alloy v1.5 upgraded to OpenTelemetry Collector v0.104.0, which included a fix to the tail sampling processor:

Previously if the decision from a policy evaluation was NotSampled or InvertNotSampled it would return a NotSampled decision regardless, effectively downgrading the result. This was breaking the documented behaviour that inverted decisions should take precedence over all others.

The “documented behavior” which the above quote is referring to is in the processor documentation :

Each policy will result in a decision, and the processor will evaluate them to make a final decision:

• When there’s an “inverted not sample” decision, the trace is not sampled;
• When there’s a “sample” decision, the trace is sampled;
• When there’s an “inverted sample” decision and no “not sample” decisions, the trace is sampled;
• In all other cases, the trace is NOT sampled

An “inverted” decision is the one made based on the “invert_match” attribute, such as the one from the string, numeric or boolean tag policy.

However, in OpenTelemetry Collector v0.116.0 this fix was reverted :

Reverts #33671, allowing for composite policies to specify inverted clauses in conjunction with other policies. This is a change bringing the previous state into place, breaking users who rely on what was introduced as part of #33671.

Breaking change: Change default value of max_streams in otelcol.processor.deltatocumulative

The default value was changed from 0 to 9223372036854775807 (max int).

Breaking change: Change default value of namespace in otelcol.connector.spanmetrics

The default value was changed from "" to "traces.span.metrics".

Breaking change: The component otelcol.exporter.logging has been removed in favor of otelcol.exporter.debug

Both components are very similar. More information can be found in the announcement issue.

Breaking change: Change default value of revision in import.git

The default value was changed from "HEAD" to "main". Setting the revision to "HEAD", "FETCH_HEAD", "ORIG_HEAD", "MERGE_HEAD" or "CHERRY_PICK_HEAD" is no longer allowed.

Breaking change: Some debug metrics for otelcol components have changed

For example, otelcol.exporter.otlp ’s exporter_sent_spans_ratio_total metric is now otelcol_exporter_sent_spans_total. You may need to change your dashboard and alert settings to reference the new metrics. Refer to each component’s documentation page for more information.

Breaking change: The convert_sum_to_gauge and convert_gauge_to_sum functions in otelcol.processor.transform change context

The convert_sum_to_gauge and convert_gauge_to_sum functions must now be used in the metric context rather than in the datapoint context. This is due to a change upstream.

Breaking change: Renamed metrics in beyla.ebpf

process.cpu.state is renamed to cpu.mode and beyla_build_info is renamed to beyla_internal_build_info.

NOTE

See the deprecations section for important information about the Loki Helm charts.

Features and enhancements

Key features in Loki 3.7.0 include the following:

• Helm charts: - Effective March 16, 2026, the Grafana Loki Helm chart has been be forked to the new repository grafana-community/helm-charts. The motivation behind this change is to accelerate development and enable direct community maintenance, as outlined in grafana/helm-charts#3955. The Grafana Community now owns maintenance of the Loki Helm Chart, while Grafana Labs continues to maintain the Helm Chart for Grafana Enterprise Logs (GEL). The Loki Helm Charts documentation has been updated to point to the new Grafana-Community repo.
• Docs: The documentation now includes a new, comprehensive Troubleshooting section, documenting the Loki and Logs Drilldown error messages.

NOTE

Note that when appropriate, documentation updates have been backported to previous releases.

Other improvements include the following:

• canary: Support passing arbitrary set of labels to use for the query (#17008) (993b3ae)
• compactor: Add support for storing chunk deletion markers in object storage instead of local disk (#19689) (856c11d)
• distributor: Add gauge to track in-flight bytes (#20091) (23ef8ec)
• distributor: Add resolved policy to blocked and enforced label error (#19826) (48d13d1)
• distributor: Add segmentation keys and resolver (#19927) (c853f2c)
• distributor: Add support for UpdateRates RPC to distributors (#19918) (9018886)
• distributor: Randomly distribute requests to the ingest-limits frontend (#19840) (1605a38)
• distributor: Shuffle shard on tenant rate limit (#19990) (3904c2b)
• distributor: Write to dataobj partitions based on segmentation key (#19946) (3a24f5d)
• helm: nameOverride now passed through helm tpl function. (#19590) (7f56fd2)
• helm: Add ability to toggle grpclb port for query frontend service (#19609) (9c4f022)
• helm: Add startupProbe to distributor (#20073) (5b76589)
• helm: Allow configuration of service trafficDistribution parameter (#19558) (55f95e3)
• helm: Allow set topologySpreadConstraints on singleBinary (#19534) (265601f)
• helm: Make loki-canary readinessProbe configurable via values.yaml (#19328) (7231766)
• helm: Use fsGroupChangePolicy=OnRootMismatch to speed up pod starts (#13942) (c7cec3a)
• index-gateway: Client side index gateway shuffle sharding (#20124) (326c7d1)
• ingest-limits: Add UpdateRates RPC, update rates from the frontend, return no-op in the service (#19894) (e173cf4)
• ingest-limits: Check partition state in parallel (#19884) (b8536aa)
• ingester: Handle state change lock in prepare downscale (#20141) (de092da)
• logcli: Allow custom headers to be passed (#20231) (c524203)
• loki: Add loki health command (#20313) (ef69cfd)
• lokitool: Add regex namespace filtering (#20209) (0c1561d)
• operator: Add option to disable ingress (#19382) (9dc71a6)
• querier: Ability to send query context for limit enforcement (#19900) (1a66d2d)

Key features in Loki 3.7.3 include the following:

• logql: Add case-insensitive regex optimization (#20578) (f6a6a4f)
• loki: Allow disabling cache for query_range requests (#21112) (1242aa5)
• operator: Add option to customize gateway server cert (#20325) (3c8b9e8)
• operator: Update metrics authentication to remove dependency on kube-rbac-proxy (#20853) (2de6e79)

Deprecations

One of the focuses of Loki 3.0 was cleaning up unused code and old features that had been previously deprecated but not removed. Loki 3.0 removed a number of previous deprecations and introduces some new deprecations. Some of the main areas with changes include:

• Promtail was deprecated in Loki 3.0 and has been removed as of Loki 3.7.3 (#21248). The code has been merged into Grafana Alloy. You can find migration documentation and a utility to convert your Promtail configuration to Alloy configuration in the Alloy documentation. Note that this removal does NOT include the lambda-promtail client.
• Deprecated storage options including the deprecation of the BoltDB store.
• Deprecated configuration options.
• API endpoint deprecations.

To learn more about breaking changes in this release, refer to the Upgrade guide.

Upgrade Considerations

For important upgrade guidance, refer to the Upgrade Guide.

• BREAKING CHANGE - engine: Make scheduler aware of total compute capacity (#19876)
• BREAKING CHANGE - engine: Share worker threads across all scheduler connections (#20229)
• BREAKING CHANGE - labels: parsed labels should not override structured metadata (#19991)
• BREAKING CHANGE - operator: Switch default for OpenShift stream labels (#21001)

Bug fixes

3.7.6 (2026-08-05)

• queryrange: Preserve sketch in MergeLabels (#23770) (d48329d)

3.7.5 (2026-08-05)

• ci: Build deb/rpm packages again in the release workflow (#23598) (6f49272)
• ci: Disable docker plugin publishing on release-3.7.x (#23440) (6095f71)
• ingester: Fix flush race in ingester (#23682) (ff152dc)
• security/UNKNOWN/: Update module github.com/klauspost/compress to v1.18.7 (release-3.7.x) (#23610) (941b557)
• security/UNKNOWN/: Update module golang.org/x/text to v0.39.0 (release-3.7.x) (#23418) (2347a9a)
• security/UNKNOWN/: Update module google.golang.org/grpc to v1.82.1 (release-3.7.x) (#23405) (f290af0)
• security/UNKNOWN/cmd/chunks-inspect: Update module github.com/klauspost/compress to v1.18.7 (release-3.7.x) (#23611) (11c4b29)
• security/UNKNOWN/cmd/dataobj-inspect: Update module github.com/klauspost/compress to v1.18.7 (release-3.7.x) (#23612) (1e3108e)
• security/UNKNOWN/cmd/dataobj-inspect: Update module go.opentelemetry.io/otel to v1.42.0 (release-3.7.x) (#23518) (28917d3)
• security/UNKNOWN/cmd/dataobj-inspect: Update module golang.org/x/text to v0.39.0 (release-3.7.x) (#23419) (5d2b9d4)
• security/UNKNOWN/operator: Update module github.com/klauspost/compress to v1.18.7 (release-3.7.x) (#23613) (8d71c59)

3.7.4 (2026-07-22)

• ci: Fix zizmor findings for operator-images in Loki 3.7 (#22820) (abc1186)
• ci: Helm CI warning fix (#22605) (7389428)
• ci: Re-enable docker plugin publishing (#22818) (28f44a8)
• compactor: Fix delete request when using Thanos objstore client with filesystem backend (#22811) (dd6fdb5)
• deps: Update module charm.land/bubbletea/v2 to v2.0.7 (release-3.7.x) (#22768) (e11f935)
• deps: Update module charm.land/lipgloss/v2 to v2.0.4 (release-3.7.x) (#22769) (10213c2)
• deps: Update module github.com/aws/aws-sdk-go-v2/credentials to v1.19.24 (release-3.7.x) (#22776) (3daf198)
• deps: Update module github.com/aws/smithy-go to v1.27.3 (release-3.7.x) (#22795) (1f6a191)
• deps: Update module github.com/baidubce/bce-sdk-go to v0.9.270 (release-3.7.x) (#22783) (103ccd8)
• deps: Update module github.com/coder/quartz to v0.3.1 (release-3.7.x) (#22784) (3509e31)
• deps: Update module github.com/IBM/ibm-cos-sdk-go to v1.14.1 (release-3.7.x) (#22785) (d0abe72)
• deps: Update module github.com/klauspost/compress to v1.18.6 (release-3.7.x) (#22797) (327da7c)
• deps: Update module github.com/pierrec/lz4/v4 to v4.1.27 (release-3.7.x) (#22798) (6b3297f)
• deps: Update module github.com/shirou/gopsutil/v4 to v4.26.5 (release-3.7.x) (#22800) (cb7f17d)
• security/CRITICAL/clients/cmd/fluentd/docker: Update dependency fluentd to v1.19.3 (release-3.7.x) (#22693) (6d8f1b5)
• security/HIGH/: Bump Go to 1.26.5 to address CVE-2026-39822 and CVE-2026-42505 (#23393) (b318f28)
• security/HIGH/: Update module github.com/apache/thrift to v0.24.0 (release-3.7.x) (#22767) (36eb5f7)
• security/UNKNOWN/: Update module github.com/containerd/containerd/v2 to v2.0.10 (release-3.7.x) (#22478) (6fa6165)
• security/UNKNOWN/pkg/push: Update module golang.org/x/net to v0.55.0 (release-3.7.x) (#22200) (dd7a124)
• storage: Fix index filename issue with legacy S3 client and chunk_delimiter (#23165) (0eca83c)

3.7.3 (2026-06-24)

• deps: Move to Go 1.26 (#20988) (4fb960c).
• deps: Update go toolchain directive to v1.25.11 (#22192) (d5d7a54).
• deps: Update golang.org/x/crypto to v0.52.0 (#22189) (6383bfe).
• deps: Update golang.org/x/net to v0.55.0 (#22190) (3d93fb8).
• deps: Update golang.org/x/sys to v0.44.0 (#22191) (3379538).
• deps: Update module github.com/containerd/containerd/v2 to v2.0.9 (#22187) (365c7e9).
• deps: Update module go.opentelemetry.io/otel to v1.41.0 (#22188) (dc8f4de).
• distributor: Sanitize User-Agent to valid UTF-8 before using as Prometheus label value (#21162) (ed3cf08).
• engine: Align range-query evaluation timestamps to the step grid (#20904) (28d03ec).
• engine: Fix additional instances of writing/reading on a closed conn (#20982) (1f7cacf).
• ingester: Duplicate lines with different structured metadata (#21039) (60a1567).
• ingester: WAL replay hang when flush doesn’t progress (#21176) (fb58ab8).
• logql: Return error on way-too-long label string (#20937) (9c9660a).
• operator: Fix incorrect image reference in OLM bundle (#20966) (84bc9bc).
• operator: Properly handle the deprecated replicationFactor (#20935) (1b96366).
• operator: Provide AWS region in environment variable for STS mode (#21136) (cb261d9).
• operator: Set ingester minimum available pods based on replication factor (#19517) (efc63e8).
• querier: Prevent goroutine leak in downstreamer on context cancellation (#21062) (2891919).
• query-engine: Fix OR operations being silently dropped (#21204) (20ea556).
• security: Mask all credentials (#20766) (f8d94cb).
• storage: Broken get and replace for S3 client (#21163) (e6abd15).

3.7.2 (2026-05-13)

• deps: CVEs in release 3.7.x (#21771) (bb4c5d8).
• deps: Update module github.com/aws/aws-sdk-go-v2/service/s3 to v1.97.3 (#21457) (7bc9450).
• ruler: Fix ruler panic related to unset validation scheme (backport release-3.7.x) (#21401) (cf65729).
• storage: Attach SHA-256 checksum on PutObject for Object Lock buckets (#21849) (7df13d9).

3.7.1 (2026-03-27)

• deps: Upgrade Go and gRPC versions on 3.7.x (#21282) (2c8fff2).

3.7.0 (2026-03-26)

• cd: Add loki-image to needs (#19870) (a2c4ea6)
• compactor: Compactor file descriptor leak (#20077) (0c3dd8c)
• config: Migrate renovate config (#19436) (97745fe)
• dataobj: Flush into multiple index objects when ErrBuilderFull (#19223) (32dbef9)
• deps: Update dataobj-inspect transitive deps version (#19813) (5b212b7)
• deps: Update module cloud.google.com/go/bigtable to v1.41.0 (main) (#20352) (6102309)
• deps: Update module cloud.google.com/go/pubsub to v2 (main) (#19803) (d47dde3)
• deps: Update module cloud.google.com/go/storage to v1.59.0 (main) (#20407) (5c71db6)
• deps: Update module github.com/alecthomas/chroma/v2 to v2.22.0 (main) (#20409) (c64f044)
• deps: Update module github.com/apache/arrow-go/v18 to v18.5.0 (main) (#20354) (d0861a1)
• deps: Update module github.com/aws/aws-sdk-go-v2 to v1.40.1 (main) (#20137) (e106809)
• deps: Update module github.com/aws/aws-sdk-go-v2/config to v1.32.7 (main) (#20401) (50ce71a)
• deps: Update module github.com/aws/aws-sdk-go-v2/credentials to v1.19.7 (main) (#20402) (f20228d)
• deps: Update module github.com/aws/aws-sdk-go-v2/service/dynamodb to v1.53.6 (main) (#20403) (31a870c)
• deps: Update module github.com/aws/aws-sdk-go-v2/service/s3 to v1.95.1 (main) (#20406) (c7c1411)
• deps: Update module github.com/aws/smithy-go to v1.24.0 (main) (#20117) (b0efa70)
• deps: Update module github.com/axiomhq/hyperloglog to v0.2.6 (main) (#20341) (4469f82)
• deps: Update module github.com/baidubce/bce-sdk-go to v0.9.256 (main) (#20342) (69e6254)
• deps: Update module github.com/bits-and-blooms/bloom/v3 to v3.7.1 (main) (#19755) (af47e1f)
• deps: Update module github.com/bmatcuk/doublestar/v4 to v4.9.2 (main) (#20418) (078dc94)
• deps: Update module github.com/coder/quartz to v0.3.0 (main) (#19786) (6f784f9)
• deps: Update module github.com/docker/docker to v28.5.2+incompatible (main) (#19756) (1007ee4)
• deps: Update module github.com/gocql/gocql to v2 (main) (#19794) (898b6d2)
• deps: Update module github.com/google/renameio/v2 to v2.0.2 (main) (#20411) (cc43074)
• deps: Update module github.com/grafana/loki/v3 to v3.6.3 (main) (#20343) (d1ae7a1)
• deps: Update module github.com/grpc-ecosystem/go-grpc-middleware/v2 to v2.3.3 (main) (#19758) (8133da9)
• deps: Update module github.com/hashicorp/consul/api to v1.33.0 (main) (#19788) (e417259)
• deps: Update module github.com/ibm/go-sdk-core/v5 to v5.21.2 (main) (#19988) (d8ab970)
• deps: Update module github.com/ibm/ibm-cos-sdk-go to v1.13.0 (main) (#20364) (52d1d8d)
• deps: Update module github.com/ibm/sarama to v1.46.3 (main) (#19760) (4a19787)
• deps: Update module github.com/influxdata/telegraf to v1.37.0 (main) (#20356) (dc1e0ae)
• deps: Update module github.com/klauspost/compress to v1.18.2 (main) (#20108) (f4f2b2a)
• deps: Update module github.com/leodido/go-syslog/v4 to v4.3.0 (main) (#19416) (036387b)
• deps: Update module github.com/minio/minio-go/v7 to v7.0.98 (main) (#20436) (cf89342)
• deps: Update module github.com/ncw/swift/v2 to v2.0.5 (main) (#19764) (fa5e144)
• deps: Update module github.com/oschwald/geoip2-golang to v2 (main) (#19799) (33eeab6)
• deps: Update module github.com/oschwald/geoip2-golang/v2 to v2.1.0 (main) (#20357) (8853d71)
• deps: Update module github.com/parquet-go/parquet-go to v0.27.0 (main) (#20426) (a283eac)
• deps: Update module github.com/prometheus/alertmanager to v0.30.0 (main) (#20358) (f53a609)
• deps: Update module github.com/prometheus/client_golang to v1.23.2 (main) (#19763) (8317f7e)
• deps: Update module github.com/prometheus/common to v0.67.5 (main) (#20363) (aaacbf4)
• deps: Update module github.com/prometheus/prometheus to v0.309.1 (main) (#20388) (bf79bcf)
• deps: Update module github.com/prometheus/sigv4 to v0.4.0 (main) (#20386) (2f80526)
• deps: Update module github.com/redis/go-redis/v9 to v9.17.2 (main) (#20116) (434a929)
• deps: Update module github.com/schollz/progressbar/v3 to v3.19.0 (main) (#20365) (0b238bc)
• deps: Update module github.com/shirou/gopsutil/v4 to v4.25.12 (main) (#20347) (0740eb8)
• deps: Update module github.com/sirupsen/logrus to v1.9.4 (main) (#20447) (35c8df7)
• deps: Update module github.com/sony/gobreaker/v2 to v2.4.0 (main) (#20366) (090ffd5)
• deps: Update module github.com/tjhop/slog-gokit to v0.1.5 (main) (#19808) (615413e)
• deps: Update module github.com/twmb/franz-go to v1.20.6 (main) (#20348) (5ee4fee)
• deps: Update module github.com/twmb/franz-go/pkg/kadm to v1.17.1 (main) (#19790) (1dad0be)
• deps: Update module github.com/twmb/franz-go/pkg/kmsg to v1.12.0 (main) (#19791) (f28c247)
• deps: Update module github.com/workiva/go-datastructures to v1.1.7 (main) (#19766) (f5e0683)
• deps: Update module github.com/xdg-go/scram to v1.2.0 (main) (#20046) (9e52320)
• deps: Update module go.opentelemetry.io/collector/pdata to v1.49.0 (main) (#20371) (7d759f2)
• deps: Update module go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc to v0.64.0 (main) (#20372) (9da1b1b)
• deps: Update module go.opentelemetry.io/contrib/instrumentation/net/http/httptrace/otelhttptrace to v0.64.0 (main) (#20373) (719635a)
• deps: Update module go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp to v0.64.0 (main) (#20374) (f1b3e1b)
• deps: Update module go.opentelemetry.io/otel/sdk to v1.39.0 (main) (#20376) (95b82d6)
• deps: Update module golang.org/x/crypto to v0.44.0 (main) (#19776) (c85c67a)
• deps: Update module golang.org/x/net to v0.47.0 (main) (#19850) (5c422a6)
• deps: Update module golang.org/x/oauth2 to v0.34.0 (main) (#20160) (ebf7b93)
• deps: Update module golang.org/x/sync to v0.19.0 (main) (#20161) (817b9d8)
• deps: Update module golang.org/x/sys to v0.40.0 (main) (#20378) (9538b20)
• deps: Update module golang.org/x/text to v0.33.0 (main) (#20408) (9fd6733)
• deps: Update module golang.org/x/time to v0.14.0 (main) (#19782) (74f68fa)
• deps: Update module google.golang.org/api to v0.257.0 (main) (#20143) (5f1da75)
• deps: Update module google.golang.org/grpc to v1.77.0 (main) (#19945) (f3213bb)
• deps: Update module google.golang.org/protobuf to v1.36.11 (main) (#20349) (a80b52e)
• deps: Update module k8s.io/apimachinery to v0.35.0 (main) (#20381) (731e067)
• distributor: Avoid recalculating the segmentation key hash twice (#19961) (8b78f79)
• distributor: Panic when no healthy instances found (#19998) (1c5dfed)
• distributor: Track discarded entries and bytes when hitting stream limits using the ingest limits service (#20244) (9b0af7c)
• docker: Missing permissions to start docker (#19947) (39d2bea)
• docker: Set WORKDIR to root in loki Dockerfiles (#19941) (13f2b1a)
• helm: Bump Helm deps, publish loki-helm-test w/ release (#19939) (7e4e34e)
• helm: Add startup probe read (#19708) (bce87fb)
• helm: Apply fix from #14126 to example (#20252) (716563a)
• helm: Correct GEL provisioner tenant creation instructions (#20271) (9639e2e)
• helm: Correct GEL version (#19657) (cd8b195)
• helm: Do not mark loki.storage.bucketNames.chunks as required, if an s3 url is provided (#19873) (e9951bb)
• helm: Do not mark loki.storage.bucketNames.chunks as required, if minio is used. (#19871) (eddd4f8)
• helm: Do not mark loki.storage.bucketNames.ruler as required, if rulerConfig.storage.type is local (#19882) (f2f564a)
• helm: Don’t fail for missing bucket name, if local disk is used. (#19675) (ad0a29e)
• helm: Don’t fail for missing bucket name, if minio is enabled. (#19745) (cd0c578)
• helm: Enable volumeAttributesClassName attributes for volumeClaimTemplates (#19719) (06da42a)
• helm: Fix ingester-b volumeAttributesClassName templating (#20188) (d696f18)
• helm: Fix rendering of dnsConfig for backend, read, write, single-binary and table-manager (#20013) (1cdb3c7)
• helm: Respect global registry in sidecar image (#18246) (#19347) (79eae2c)
• helm: Update Chart version in README.md (#19669) (94096b7)
• helm: Update version (#19670) (b90ae22)
• ingest-limits: Reduce FetchMaxBytes to 10MB (#19883) (82cfafd)
• ingest-limits: Revoke partitions if lost (#20030) (1ac5d1f)
• kafka: Deadlock on shutdown (#20384) (272a278)
• logql: Parsed labels should not override structured metadata (#19991) (61f9367)
• lokitool: Update ruler path and enable alternative TLS env variables (#19572) (d1ce5cb)
• nomad: Nomad simple example (#19629) (17aec11)
• operator: Change leader-election parameters (#19707) (86068cf)
• operator: Do not deploy NetworkPolicies automatically on OCP 4.20 (#19680) (8df33ff)
• operator: Return quickstart script to working condition and improve rootless usage (#19960) (397da27)
• parser: Do not cache key conflicts results in intern set (#19984) (0a9b024)
• querier: Support multi-tenant queries in Patterns API (#19809) (f609e27)
• querier: Restrict start/end timestamp to requested range for scheduler (#20086) (448cc74)
• querier: Set Content-Type header for JSON responses in serializeHTTPHandler (#19878) (019d6b4)
• querylimits: Accept request limits over not initialized limits (#19891) (905eac8)
• retry: Do not retry if any of multierrors is a client error (#19887) (9825137)
• ruler: Validate remote write config (#19920) (e916944)
• server: Return status bad request code for interval limit error (#19895) (f21f5d5)
• storage: Do not override S3 region if already specified in configuration chain (#20127) (0046bfb)
• storage: Fix regression in S3 client configuration (#20110) (d3f9532), closes #19908

NOTE

See the deprecations section for important information about the Loki Helm charts.

Features and enhancements

Key features in Loki 3.6.0 include the following:

• Horizontally scalable compactor: The Compactor in Loki is a singleton service that compacts the index and applies retention, including processing delete requests. Because it can only be scaled vertically, it limits how many resources you can add to speed things up when the workload increases in bigger cells. This release adds a horizontally scalable compactor, to speed up processing large delete requests or catching up if there has been a lag in processing for some reason. Note that the horizontally scalable compactor delegates deletion work to the workers. Loki continues to perform compaction and apply retention in the main Compactor. For more information, refer to the documentation.
• OTel: Loki has been refactored to use the OpenTelemetry tracing library instead of the OpenTracing library. (#17859). Note that Loki can still be configured using JAEGER_ env vars and still exports traces in JAEGER_ format.
• meta-monitoring: Moves the meta-monitoring responsibilities from the Grafana meta-monitoring Helm chart to the widely supported Grafana Kubernetes Monitoring Helm chart. This was implemented in two PRs: Meta Monitoring Loki v2 (Part 1) - Telemetry Collection (#16665) (2d0bbfd) and Meta Monitoring Loki v2 (Part 2) - Loki Mixin’s (#16962) (93360f6).
• Helm charts: The Grafana Developer Advocates launched a pilot Community Helm Maintainers group at the end of June made up of Grafana Champions who expressed interest in helping maintain the Loki Helm chart. We have resumed regular releases of the Loki Helm chart and reduced the backlog of open Loki Helm chart pull requests from around 150 to less than 40. For information about recent changes, refer to the Helm chart CHANGELOG.
• UI: We have moved the Loki UI to a Grafana Plugin. The static Operational UI is not going away, and all the APIs that serve it will remain in Loki. The JavaScript that serves the UI, however, has been moved to the Loki Operational UI Grafana plugin (#19390).
• Docs: Published a third Cloud Deployment Guide, this one for Google Cloud Platform (#17503). Added links to three log related Learning paths to the Loki docs (#17956). Revised the LogQL query page (#16286) and added a LogQL video (#18111). Added a caution to the Labels topic and added a new topic with instructions how to modify the default OpenTelemetry resource attributes that Loki converts to labels (#17958). Updated the caching topic (#17420).

NOTE

Note that when appropriate, documentation updates have been backported to previous releases.

Other improvements include the following:

• analytics: Change how analytics proxy works (#19098)
• API: Add endpoint returning tenant configured limits (#17101) (6818c90)
• API: Loki Improvement Document (LID) for query splitting (#16958) (66f0f0b)
• compactor HS: Add a dashboard for monitoring deletion with horizontally scalable compactor (#18588) (10db8ab)
• compactor HS: Add a job runner for processing of deletion jobs (#18058)
• compactor HS: Add job queue and deletion job builder (#17843)
• compactor HS: Add jsonnet for deploying horizontally scalable compactor (#18550) (5e28950)
• compactor HS: Add support for applying storage updates after finishing processing of deletion manifest (#18294) (ba3c670)
• compactor HS: Add support for worker for processing of jobs from the compactor’s job queue (#18165) (d05c4bc)
• compactor HS: Fixes and changes for getting horizontally scalable compactor to work well (#18526) (000f90d)
• compactor HS: Refactor interfaces and build a link between delete requests manager and compactor for HS of compactor (#18326) (3646607)
• compactor HS: Store index updates per processed segment in the manifest to the object storage (#18212) (8007e22)
• compactor HS: Tune compactor worker config and add experimental feature warning to the docs (#18539) (5bf3f74)
• compactor HS: Use streaming RPCs for interacting with job queue over grpc (#18125) (385e317)
• compactor HS: Wire things up and add some metrics (#18421) (bd29320)
• compactor HS: Add a metric to track the number of jobs left to conclude ongoing unit of work (#19506)
• distributor: Add MaxRecvMsgSize config for uncompressed message size limits (#16915) (1d99f4d)
• distributor: Log ingestion policy when log_push_request is set (#18712) (0ec787f)
• distributor: Mark truncated log lines with identifier (#18262) (0ee8e76)
• distributor: Stop pushing to the metadata topic from distributors (#17525) (e0f015c)
• distributor: Remove tenant topic tee (#19082)
• fluentbit: Update fluent/fluent-bit to v4 which adds support for structured metadata to the out_loki plugin (#17200) (f108e1c)
• fluentd: Add ‘compress gzip’ setting to fluentd plugin (#18995)
• helm: Add block_builder config to helm chart (#17451) (a9808a1)
• helm: Add startupProbe to memcached(#19166)
• helm: Add startupProbe to memcached exporter(#19165)
• helm: Add configurable livenessProbe to single binary StatefulSet(#19072)
• helm: Add dnsconfig to all missing resources(#18903)
• helm: Add FOLDER_ANNOTATION logic for sidecar container in loki helm chart, making multi-tenant alerting easier (#13289) (833fcbc)
• helm: Add health probes to admin-api (#18546) (358228a)
• helm: Add health probes to bloom-builder (#18549)
• helm: Add health probes to bloom-gateway(#18551)
• helm: Add health probes to bloom-planner(#18552)
• helm: Add health probes to gateway component(#18545)
• helm: Add health probes to loki in backend(#18548)
• helm: Add location snippet to nginx config (#18105) (5737751)
• helm: Add missing revisionHistoryLimit (#18638) (d0e0d1a)
• helm: Add more specificity to affinity and topology spread constraints(#19061)
• helm: Add namespaceOverride(#18707)
• helm: Add non-default-values.yaml to increase test coverage (#18360) (b0e04f4)
• helm: Add option to use deployment type instead of daemonset in canary (#17364) (a6563cf)
• helm: Add PDB for patternIngester (#17058) (7fad38d)
• helm: Add readiness and liveness support for all memcached components (#18525) (9b54d9f)
• helm: Add readiness probe for memcached (#15609) (6ed1461)
• helm: Add replicas to loki-canary deployment (#19095)
• helm: Add startupProbe to loki-sc-rules (#18547) (121e84e)
• helm: Add support for custom alert rule annotations (#19193)
• helm: Add tpl support for pattern_ingester and ingester_client (#16759) (901b2b0)
• helm: Added trafficDistribution for service-distributor (#17742) (78f154f)
• helm: Adds a flag for tokengen rbac components so that they can be disabled optionally (#15882) (10536b2)
• helm: Allow configuring service types (#18708)
• helm: Allow enabling user namespaces (#18661)
• helm: Allow external memcached setup (#17432) (013aa17)
• helm: Allow extraObjects to be defined as either a list or dictionary.(#13252)
• helm: Allow passing tenant password hash instead of password (#17049) (20b97f7)
• helm: Allow setting labels for ingesters (#18536) (ab426be)
• helm: Enable rules sidecar for the ruler pods (#17231) (2c5ad02)
• helm: Expose all storage configuration (#18509) (db5e021)
• helm: Expose labels on volumeClaimTemplates (#13317) (cdefbbf)
• helm: Expose Loki UI in pure ingress (#18400) (21f3d09)
• helm: Helm issue templates (#18693) (e0ec815)
• helm: Implementing support of l2 chunksCache memcache (#17556) (039e7c1)
• helm: Integrate the enterprise-provisioner into Loki helm chart (#17616)
• helm: Make access modes for persistence on all PVCs and StatefulSets editable (#13474)
• helm: Only delete PVC on statefulset deletion, not scaling (#19097)
• helm: Support bypass all s3/gcs/azure config in object_store (#17597) (d6c2a8e)
• helm: Support separate loki ruler storage configuration (#18510)
• helm: Support templated loki.operational_config (#17045) (078f353)
• helm: Update Grafana Loki version to 3.5.3 (#18619) (907bb61)
• helm: Update Loki version in charts (#18447) (e2770ae)
• helm: Use named ports for memcached services targetports (#18990)
• helm: Delete statefulset if volumeClaimTemplates changes (#19217)
• helm: Added affinity property to the loki-canary DaemonSet (#17720)
• ingester: Add ingester_chunks_flush_requests_total (#18696)
• ingester: Add per policy overridable stream limits (#18994)
• kafka: Support multiple Kafka clients in component (#17757)
• kafka: Adds partition-ingester push latency histogram (#17385) (01cf36d)
• kafka: Optionally enable Kafka histograms for read/write latency (#17089) (e962398)
• kafka: Refactors kafka/client to improve metric collection (#17308) (ab65ac5)
• kafka: Support consuming records from Kafka (#17691)
• limits: Add ability to filter tenant limits endpoint by allowlist (#18521) (73de7b5)
• logcli: Add delete commands to logcli (#18375) (f38950a)
• logcli: Add --log.level parameter to help with diagnosing errors when using logcli. (#19136)
• logs drilldown: Return metric query partial results for drilldown (#18286) (c0c2265)
• logs drilldown: Drilldown config endpoint (#19028)
• logs drilldown: Implement unwrap as a projection (#19409)
• logs drilldown: Persist patterns as aggregated metrics (#17737)
• logs drilldown: Query persisted patterns (#17980)
• loki-canary: Update loki canary to support batched log push (#17558)
• metrics: Adds distributor lag counter to push.go (#18012)
• metrics: Adds User-Agent to distributor push logging (#17764)
• mixins: Allow disabling Thanos dashboard and fix dashboard’s file name (#18749)
• operator: Add alert LokiIngesterFlushFailureRateCritical (#18698)
• operator: Add warning alert for when LokiStack is not getting ready (#19258)
• operator: Allow users to configure virtual-host-style for S3 secrets (#17363) (c4dace5)
• operator: Allow using resource requests as limits (#17484) (0b2bae8)
• operator: Deploy network policies with LokiStack (#19099)
• operator: Extend LokiStack authorization to support OTel Semantics (#16031)
• operator: S3 validation reject endpoints that contain a URL path (#19356)
• operator: Update Loki operand to v3.5.5 (#19187)
• pattern ingester: Pattern ingester pushes detected level as structured metadata (#16924) (66033bb)
• pattern ingester: Implement volume-based pattern filtering (#18838)
• policies: Default ingestion policy mappings merged with per-tenant mappings (#18926)
• thanos: Allow dashes in storage_prefix config (#16934) (c01d9c7)

Grafana Labs and the Loki team are excited to announce the release of Loki 3.5. Here’s a summary of new enhancements and important fixes.

For a full list of all changes and fixes, refer to the CHANGELOG.

Features and enhancements

Key features in Loki 3.5.0 include the following:

• Helm charts: Loki 3.5 includes multiple improvements to the Helm charts, including: Adding support for additional global variables in the Helm chart (#16062) (e5f31b3), adding the extraContainers property to the backend Loki pods (#16205) (528a1c6) [closes #16164], adding tpl() support back to read, write and backend pods (#16664) (af05c38), and allowing provisioner to be namespaced (#16091) (af85ecf).
• storage: Add support for `SQLite’ as a database for storing delete requests (#16437) (5b33e65) and when using SQLite, use the stored completion time to reduce the requests considered for querytime filtering (#16600) (3277321).
• Docs: This release includes a heavily revised meta-monitoring section to support new monitoring requirements for Grafana Enterprise Logs customers.

Other improvements include the following:

• distributor: Add simulated latency (#16733) (7119f6e)
• distributor: Do not enforce labels vs agg metric stream (#16696) (25c0620)
• distributor: Enforce limits in distributors (#16705) (bc2111e)
• distributor: Update log level discovery from json to detect nested fields (#16026) (86dc47b)
• distributor: Use stream labels instead of hash in rate limit reasons (#16738) (fdf8632)
• distributors: Add dry-run mode for checking limits in distributors (#16754) (d6c499b)
• kafka: tenant topics (#15977) (c258419)
• ksonnet: Assert for bad max_query_lookback configuration (#16362) (073c94c)
• ksonnet: Add no_schedule_constraints option to distributors (#16029) (9e9a09d)
• lambda-promtail: Improve relabel configuration parsing and testing (#16100) (2587f34)
• logical planning: Planner playground (#16403) (89f0ed7)
• logql: Add LogQL Benchmark Suite (#16434) (1b57297)
• operator: Add configuration option for dropping OTLP attributes (#15857) (bd1ea23)
• operator: Add support for Swift TLS CA configuration (#15260) (62a72f6)
• operator: Enable time-based stream-sharding (#16390) (1b4f1f5)
• operator: Update Loki operand to v3.4.2 (#16360) (42f87d3)
• storage: Add S3 chunk delimiter configuration to support MinIO running on Windows (#16319) (8429b0f)
• storage: Add a configuration option for custom GCS endpoints (#16419) (f1ebd97)

Deprecations

One of the focuses of Loki 3.0 was cleaning up unused code and old features that had been previously deprecated but not removed. Loki 3.0 removed a number of previous deprecations and introduces some new deprecations. Some of the main areas with changes include:

• Promtail is deprecated, as the code has been merged into Grafana Alloy. You can find migration documentation and a utility to convert your Promtail configuration to Alloy configuration in the Alloy documentation. Note that this deprecation does NOT include the lambda-promtail client.
• Deprecated storage options including the deprecation of the BoltDB store.
• Deprecated configuration options.
• API endpoint deprecations.

To learn more about breaking changes in this release, refer to the Upgrade guide.

CAUTION

Grafana Alloy is the new name for our distribution of the OTel collector. Grafana Agent has been deprecated and is in Long-Term Support (LTS) through October 31, 2025. Grafana Agent will reach an End-of-Life (EOL) on November 1, 2025. Read more about why we recommend migrating to Grafana Alloy.

Upgrade Considerations

For important upgrade guidance, refer to the Upgrade Guide.

• BREAKING CHANGE - ksonnet: Clean-up deprecated configurations (#16007).
• BREAKING CHANGE - operator: Add configuration option for dropping OTLP attributes (#15857).

Bug fixes

[Followed by detailed list of bug fixes and dependency updates for versions 3.5.12 down to 3.5]

Features and enhancements

Grafana Labs and the Loki team are excited to announce the release of Loki 3.2. Explore Logs is also now Generally Available. Upgrade to Loki/GEL 3.2 to get the best possible experience with Explore Logs. Here’s a summary of new enhancements and important fixes.

For a full list of all changes and fixes, refer to the CHANGELOG.

Key features in Loki 3.2.0 include the following:

• API: The Loki APIs have been updated to fail log queries when executed on instant query endpoint (#13421) (ce71f1c), add structured metadata to /detected_fields API (#13604) (ce02cc2), and avoid looking up unnecessary TSDB symbols during Volume API (#13960) (7c1a849).
• Query acceleration with Bloom filters (experimental): The Bloom filters feature introduced in Loki 3.0 remains experimental. Note that the Bloom Compactor component has been removed (#13969) (b75eacc) and replaced by the Bloom Planner and Bloom Builder components (#13997) (be9eb50).
• distributor: Ignore empty streams in distributor if all entries fail validation (#13674) (6c4b062), and limit to block ingestion until configured date (#13958) (b5ac6a0).
• Explore Logs Is now Generally Available (GA). For the best experience, you should be on Grafana 11.2 or later and Loki 3.2. This release includes enhancements to add _extracted suffix to detected fields conflicts (#13993) (ab1caea), collect and serve pre-aggregated bytes and counts (#13020) (467eb1b), and remove cardinality filter (#13652) (4f534d7).
• Helm: This release includes updates to the Helm charts to make gateway container port configurable. (#13294) (05176e4) and to support alibabacloud oss in the Helm chart (#13441) (3ebab6f). It also includes a breaking change to the Helm chart to support distributed mode and 3.0 (#12067).
• ingester: Ingester Stream Limit Improvements: Ingester stream limits now take into account “owned streams” and periodically update when the Ingester ring is changed. Non-owned streams are now also flushed when this update takes place. The stream limit calculation has also been updated for improved accuracy in multi-zone ingester deployments. (#13532) (ec34aaa).
• lambda-promtail: Add S3 log parser support for AWS GuardDuty (#13148) (2d92fff), build lambda with zip file (#13787) (9bf08f7), and ensure messages to Kinesis are usable by refactoring parsing of KinesisEvent to match parsing of CWEvents + code cleanup (#13098) (dbfb19b).
• loki: Add ability to disable AWS S3 dualstack endpoints usage(#13785) (bb257f5), not enforce max-query-bytes-read and max-querier-bytes-read in limited tripperware (#13406) (47f6ea5), and upgrade Prometheus (#13671) (b88583d).
• operator: Add alert for discarded samples (#13512) (5f2a02f), add support for the volume API (#13369) (d451e23), enable leader-election (#13760) (1ba4bff), and update Loki operand to v3.1.0 (#13422) (cf5f52d).
• storage: Convert WalSegmentWriter to io.ReadSeeker (#13340) (19c0509) and instrument failed chunk encoding/decoding (#13684) (5a87ccb).
• docs: Add getting started video for ingesting OTel logs (#13226) (5e560f9).

Other improvements include the following:

• chunks-inspect: Support structured metadata (#11506) (1834065).
• exporter: Include boolean values in limit exporter (#13466) (4220737).
• mempool: Replace sync.Mutex with sync.Once (#13293) (61a9854).
• metrics: Collect duplicate log line metrics (#13084) (40ee766).

Deprecations

One of the focuses of Loki 3.0 was cleaning up unused code and old features that had been previously deprecated but not removed. Loki 3.0 removed a number of previous deprecations and introduces some new deprecations. Some of the main areas with changes include:

• Deprecated storage options including the deprecation of the BoltDB store.
• Deprecated configuration options.
• API endpoint deprecations.

To learn more about breaking changes in this release, refer to the Upgrade guide.

CAUTION

Grafana Alloy is the new name for our distribution of the OTel collector. Grafana Agent has been deprecated and is in Long-Term Support (LTS) through October 31, 2025. Grafana Agent will reach an End-of-Life (EOL) on November 1, 2025. Read more about why we recommend migrating to Grafana Alloy.

Upgrade Considerations

For important upgrade guidance, refer to the Upgrade Guide and the separate Helm Upgrade Guide.

• BREAKING CHANGE - API: Fail log queries when executed on instant query endpoint (#13421).
• BREAKING CHANGE - blooms: Remove bloom compactor component (#13969).
• BREAKING CHANGE - docker: Remove wget from Promtail docker image(#15101).
• BREAKING CHANGE - Helm: Update Helm chart to support distributed mode and 3.0 (#12067).
• BREAKING CHANGE - Helm: Fix how we set imagePullSecrets for enterprise-gateway and admin-api. (#13761) (3be5a45).
• BREAKING CHANGE - jsonnet: Convert read statefulset into deployment for loki-simple-scalable (#13977).

IMPORTANT

Out of an abundance of caution, we advise that users with Loki or Grafana Enterprise Logs (GEL) deployments on AWS upgrade their Helm charts or change the names of their buckets, as outlined in the solutions and mitigations section of this blog post.

Bug fixes

3.2.2 (2024-12-04)

• BREAKING CHANGE - docker: Remove wget from Promtail docker image(#15101).
• docker: Move from base-nossl to static. This PR removes the inclusion of glibc into most of the Docker images created by the Loki build system. (#15203).
• logql: Updated JSONExpressionParser not to unescape extracted values if it is JSON object. (#14499).
• promtail: Switch Promtail base image from Debian to Ubuntu to fix critical security issues (#15195).
• storage: Have GetObject check for canceled context. S3ObjectClient.GetObject incorrectly returned nil, 0, nil when the provided context is already canceled (#14420).

3.2.1 (2024-10-17)

• config: Copy Alibaba and IBM object storage configuration from common (#14315).
• logql: updated JSONExpressionParser not to unescape extracted values if it is JSON object. (backport release-3.2.x) (#14502).
• promtail: Revert build image to Debian Bullseye to fix libc version issue in Promtail (#14386).
• storage: have GetObject check for canceled context (backport release-3.2.x) (#14422).

3.2.0 (2024-09-19)

Extensive list of fixes and improvements related to blooms, ci, deps, detected fields, Helm, index gateways, ingester, LogQL, log results cache, loki, loki-mixin, operator, querier, query engine, retry, ruler, sharding, storage, structured metadata, WAL, and more as detailed in the full changelog.

Grafana Labs is excited to announce the release of Loki 2.9.0 Here’s a summary of new enhancements and important fixes.

For a full list of all changes and fixes, refer to the CHANGELOG.

Features and enhancements

• Structured metadata: The Structured Metadata feature, which was introduced as experimental in release 2.9.0, is generally available as of release 2.9.4.
• Query Language Improvements: Several improvements to the query language that speed up line parsing and regex matching.
• Remote rule evaluation: Rule evaluation can now be handled by queriers to improve speed.
• Multi-store Index support: Loki now supports reading/writing indexes to multiple object stores which enables the use of different storage buckets across periods for storing index.
• New volume and volume_range endpoints: Two new endpoints, index/volume and index/volume_range, have been added to Loki. They return aggregate volume information from the TSDB index for all streams matching a provided stream selector.
• New Storage Client: Add support for IBM cloud object storage as storage client.
• Block queries by hash: Queries can now be blocked by a query hash.
• logfmt stage improvements: logfmt parser now performs non-strict parsing by default which helps scan semi-structured log lines.

Deprecations

• Legacy index and chunk stores that are not “single store” (such as tsdb, boltdb-shipper) are deprecated. These storage backends are Cassandra (cassandra), DynamoDB (aws, aws-dynamo), BigTable (bigtable, bigtable-hashed), GCP (gcp, gcp-columnkey), and gRPC (grpc). See Configure storage for more information.
• The table-manager target is deprecated, because it is not used by “single store” implementations.
• The -boltdb.shipper.compactor.* CLI flags are deprecated in favor of -compactor.*.
• The -ingester.unordered-writes CLI flag is deprecated and will always default to true in the next major release.
• For the full list of deprecations, see CHANGELOG.md

Bug fixes

For a full list of all changes and fixes, refer to the CHANGELOG.

Grafana Labs is excited to announce the release of Loki 2.8. Here’s a summary of new enhancements and important fixes:

Features and enhancements

• TSDB index no longer experimental After extensive testing of the new TSDB index in our Grafana Cloud Logs offering, we can safely announce that it is no longer experimental and encourage all Loki deployments to use it!
• Query blocker Queries can now be blocked in the Querier/Ruler via a per-tenant runtime configuration.
• New backend target A new, third target was added to Loki’s scalable configuration, which is the default configuration used in the Loki helm chart. This allows Loki to be run as 3 targets (read, write, and backend) and makes the read target stateless and therefore able to be run as a Kubernetes deployment that can be scaled automatically.

For a full list of all changes, look at the CHANGELOG.

Upgrade Considerations

As always, please read the upgrade guide before upgrading Loki.

Bug fixes

2.8.6 (2023-10-17)

• Upgrade go to v1.20.10, golang.org/x/net to v0.17.0 and grpc-go to v1.56.3 to patch CVE-2023-39325 / CVE-2023-44487

2.8.5 (2023-09-14)

• Update Docker base images to mitigate security vulnerability CVE-2022-48174

2.8.2 (2023-05-03)

• Update Go to 1.20.4 to address security vulnerabilities in the previous Go version.
• Promtail: Add new decompression configuration to customize the decompressor behavior.

2.8.1 (2023-04-21)

• Fix bug that was dropping index if period is a zero value.
• Fix redis client to prevent it from incorrectly choosing cluster mode with local address.
• Update go to 1.20.3 to address security vulnerabilities in the previous go version.
• Update alpine image to 3.16.5 to address security vulnerabilities in the previous version.
• Promtail: Fix journald support in amd64 binary build.

For a full list of all changes and fixes, look at the CHANGELOG.