Tags

3.3.0

- [metrics] Remove some SM/PM for external components that should not be in this chart.
- [metrics] Allow to optionally install api_server and kubelet SMs.
- [docs] mention Fluent Bit memory limits configuration.
- [metrics] Fix regression in on-disk store configuration.

3.2.0

- [logs] Allow adding filters at the beginning of the logs pipeline.
- [all] Define all images in values.yaml to support TN installation.
- [prometheus] ConfigReloader: default to operator's tag if not provided.
- [all/fluent-bit] Revert "Pull stream pristine image via registry.cern.ch" do to TN limitations.

3.1.0

- [logs] Allow adding custom filters for log parsing/massaging.
- Add extra ignored mount points for cvmfs and eos to Node Exporter.
- [logs] [events] Homogenise label name containing the Kubernetes cluster name.
- [logs] Expose all Kubernetes prefixed by kubernetes_.
- [docs] Recommend a given master flavor.

3.0.0

Although there hasn't been any breaking changes in the chart itself,
we're bumping the major as there has been a fluent-bit major bump.

- Upgrade fluent-bit to 4.0.1.
- Roll metrics' statefulset automatically upon config changes.
- Pull stream pristine image via registry.cern.ch.
- Add option to only send logs from a list of namespaces.
- Roll daemonset automatically upon config changes.
- Minor documentation tweaks.

2.1.0

This is a maintenance release which does not contain many user-visible
bugfixes or features, focusing on refinements in the tooling, the
development workflows and the documentation (not mentioned in the list
below -- see Git log for details).

- Support forwarding Kubernetes events to MONIT.
- Unify Fluent Bit image configuration, allowing per-component overrides
- Honour scrapping interval configuration

2.0.0

This is a major backwards incompatible release, hence the major
version bump. Please note that starting from this version there's no
need to ship the Prometheus CRDs separately (see installation
instructions for details).

- Upgrade Fluent Bit to 3.2.6.
- Ship Prometheus operator CRDs within this chart.
- Fail early if tenant.name and tenant.password are necessary and are not provided.
- Fail early if kubernetes.ClusterName are necessary and are not provided.
- Do not try to push metrics from Prometheus to local Fluent Bit if not installed.
- Don't deploy it-monit-metrics-fluentbit service if no Fluent Bit is avail.
- Fix regression not allowing to configure Alertmanager in Prometheus.
- Allow configuring mounts in Alertmanager.
- Correct Fluent Bit's extraVolumes' indentation.
- Correct Fluent Bit's extraVolumeMounts' indentation.
- Correct typos and wording in documentation.

+ Upgrade Prometheus server to 2.53.3

Includes:
+ Fix Prometheus server configuration by changing some keys from snake case to camel case.

This release includes:
+ Enable Prometheus remote write to local Fluent Bit.
+ Enable node exporter netdev collector.

Applied hot fix from f8520328:
  - hotfix: alermanager and pushgateway ingresses.

This tag includes the following changes:

- [FIX] (ebfc8594eb8) fluentbit: Conditional creation of ConfigMap :: Nacho Barrientos :: Skips ConfigMap creation if no logs are collected, matching conditions in fluentbit-logs/daemonset.yaml.

- [FEATURE] (5e18e56e8d6) push-gateway: Added component and configuration :: Guillermo Facundo Colunga :: Introduced push gateway component with deployment, service, ingress, and ServiceMonitor for metric scraping.

- [FEATURE] (8f9ca2d69a0) prometheus: Enable listening to Prometheus rules from all namespaces :: Guillermo Facundo Colunga :: Configured Prometheus to capture PrometheusRule objects across all namespaces by setting selectors to null.

- [FEATURE] (c60847c59a3) alertmanager: Inclusion in the chart :: Guillermo Facundo Colunga :: Added option to enable Alertmanager. Deployable with or without ingress, in HA or non-HA mode. Default setting is disabled.