v1.5.0 - 2023-02-03

Release Notes

New Metrics Endpoint

As of v1.5.0, we are introducing a new metrics endpoint which will replace the [Metrics Collector for Apache Cassandra (MCAC)](Metrics Collector for Apache Cassandra) in an upcoming release and is built directly in the Management API.
MCAC's architecture is not well suited for Kubernetes and the presence of collectd was both creating bugs and adding maintenance complexity.
MCAC is still enabled by default in v1.5.0 and can be disabled by setting .spec.cassandra.telemetry.mcac.enabled to false. This will disable the MCAC agent and modify the service monitors/vector config to point to the new metrics endpoint.

Note that the new metrics endpoint uses names for the metrics which are much closer to the names of the Cassandra metrics. For example, Client Requests latencies for the LOCAL_ONE consistency level will be found in the org_apache_cassandra_metrics_client_request_latency_read_local_one metric.
Another example, the SSTables per read percentile metrics for the system_traces keyspace can be found here:

org_apache_cassandra_metrics_keyspace_ss_tables_per_read_histogram_system_traces{host="0782cc86-ca60-47ac-a513-620a44c62fe4",instance="172.24.0.4",cluster="test",datacenter="dc1",rack="default",quantile="0.5",} 0.0
org_apache_cassandra_metrics_keyspace_ss_tables_per_read_histogram_system_traces{host="0782cc86-ca60-47ac-a513-620a44c62fe4",instance="172.24.0.4",cluster="test",datacenter="dc1",rack="default",quantile="0.75",} 0.0
org_apache_cassandra_metrics_keyspace_ss_tables_per_read_histogram_system_traces{host="0782cc86-ca60-47ac-a513-620a44c62fe4",instance="172.24.0.4",cluster="test",datacenter="dc1",rack="default",quantile="0.95",} 0.0
org_apache_cassandra_metrics_keyspace_ss_tables_per_read_histogram_system_traces{host="0782cc86-ca60-47ac-a513-620a44c62fe4",instance="172.24.0.4",cluster="test",datacenter="dc1",rack="default",quantile="0.98",} 0.0
org_apache_cassandra_metrics_keyspace_ss_tables_per_read_histogram_system_traces{host="0782cc86-ca60-47ac-a513-620a44c62fe4",instance="172.24.0.4",cluster="test",datacenter="dc1",rack="default",quantile="0.99",} 0.0
org_apache_cassandra_metrics_keyspace_ss_tables_per_read_histogram_system_traces{host="0782cc86-ca60-47ac-a513-620a44c62fe4",instance="172.24.0.4",cluster="test",datacenter="dc1",rack="default",quantile="0.999",} 0.0

Existing dashboards need to be updated to use the new metrics names. Note that the new metrics endpoint is always enabled, even when MCAC is enabled. This allows accessing the new metrics on port 9000 before disabling MCAC.

MCAC will be fully removed in a future release, and we highly recommend to experiment with the new metrics endpoint and prepare the switch as soon as possible.

Changelog

• [FEATURE] #783 Allow disabling MCAC
• [FEATURE] #739 Add API for cluster-level tasks
• [FEATURE] #775 Add the ability to inject and configure a Vector agent sidecar in the Cassandra pods
• [FEATURE] #600 Disable secrets management and replication with the external secrets provider
• [FEATURE] #501 Allow configuring annotations and labels on services, statefulsets, deployments and pods
• [FEATURE] #790 Deploy Vector agent to scrape Stargate metrics
• [FEATURE] #789 Deploy Vector agent for Reaper metrics scraping
• [ENHANCEMENT] Upgrade cass-operator to v1.14.0
• [ENHANCEMENT] #817 Allow configuring the Vector agent sidecar in the CRD
• [ENHANCEMENT] #796 Enable smart token allocation by default for DSE
• [ENHANCEMENT] #765 Add GitHub workflow to test various k8s versions
• [ENHANCEMENT] #323 Use Cassandra internals for JMX authentication
• [ENHANCEMENT] #770 Update to Go 1.19 and Operator SDK 1.25.2
• [ENHANCEMENT] #525 Deep-merge cluster- and dc-level templates
• [ENHANCEMENT] #781 Expose perNodeConfigInitContainer.Image through CRD
• [ENHANCEMENT] #783 Make ServiceAccount of Cassandra pods configurable
• [ENHANCEMENT] #822 Remove relabelling rules from service monitors when MCAC is disabled
• [BUGFIX] #844 Point Vector to the right path when the new metrics endpoint is used
• [BUGFIX] #726 Don't propagate Cassandra tolerations to the Stargate deployment
• [BUGFIX] #778 Fix Stargate deployments on k8s clusters using a custom domain name
• [TESTING] Pin Kustomize to v4 in CI
• [TESTING] #761 Stabilize integration and e2e tests
• [TESTING] #799 Fix GHA CI failures with Kustomize install being rate limited