Releases: sassoftware/viya4-monitoring-kubernetes
Releases · sassoftware/viya4-monitoring-kubernetes
1.0.12 - October 18, 2021
Version 1.0.12 (18OCT21)
-
Overall
- [CHANGE] The minimum supported version of OpenShift is now 4.7. OpenShift
support itself is still experimental. - [FIX] There is now a check for the presence of the
sha256sum
utility
in thePATH
- [FIX] There is now a timeout (default 10 min) when deleting namespaces
usingLOG_DELETE_NAMESPACE_ON_REMOVE
orMON_DELETE_NAMESPACE_ON_REMOVE
.
The timeout can be set viaKUBE_NAMESPACE_DELETE_TIMEOUT
.
- [CHANGE] The minimum supported version of OpenShift is now 4.7. OpenShift
-
Monitoring
- [FIX] Metrics will be properly collected from the SAS Deployment Operator
- [CHANGE] Internal improvements, refactoring and preparations for future support
of application multi-tenancy in SAS Viya - [FIX] The two SAS Jobs dashboards have been updated and slightly optimized
-
Logging
- [CHANGE] Fluent Bit has been upgraded to version 1.8.7
- [CHANGE] Internal improvements, refactoring and preparations for future support
of application multi-tenancy in SAS Viya
1.0.11 - September 13, 2021
Version 1.0.11 (13SEP21)
-
Monitoring
- [FEATURE] SAS Job dashboards now support a 'queue' filter for SAS Workload
Orchestrator - [FEATURE] SAS Job dashboards 'Job' filter now displays user-provided
job names if available - [DEPRECATION] In the next release, NodePorts will be disabled by default
for Prometheus and AlertManager for security reasons. Set the environment
variablePROM_NODEPORT_ENABLE=true
to maintain current behavior as it will
default to 'false' in the next release
- [FEATURE] SAS Job dashboards now support a 'queue' filter for SAS Workload
-
Logging
- Internal improvements, refactoring and preparations for application multi-
tenancy in SAS Viya
- Internal improvements, refactoring and preparations for application multi-
1.0.10 - August 16, 2021
Version 1.0.10 (16AUG21)
-
Overall
- [FEATURE] The version of
viya4-monitoring-kubernetes
deployed is now
saved in-cluster for support purposes
- [FEATURE] The version of
-
Monitoring
- [FIX] Grafana update fails with PVC multi-attach error
-
Logging
- [FEATURE] SAS Update Checker Report added to Kibana
- [FIX] Enabled NodePort for Elasticsearch causes update-in-place to fail
- [FIX] Eventrouter references deprecated version of K8s authorization API
1.0.9 - July 19, 2021
Version 1.0.9 (19JUL21)
-
Overall
- [FEATURE] OpenShift version checking has been added
- Version 4.6.x is supported
- Version 4.5 or lower generates an error
- Version 4.7 or higher generates a warning
- [FEATURE] Integration with the SAS Viya workload node placement strategy
can be enabled withNODE_PLACEMENT_ENABLE=true
- [FEATURE] OpenShift: Path-based ingress can be enabled
withOPENSHIFT_PATH_ROUTES=true
- [FEATURE] OpenShift version checking has been added
-
Monitoring
- [FIX] OpenShift: Some of the Perf dashboards displayed empty charts
- [CHANGE] Prometheus version changed from 2.26.0 to 2.26.1
- [FEATURE] OpenShift: A custom route hostname can be set with
OPENSHIFT_ROUTE_HOST_GRAFANA
- [FIX] The memory limit of the Prometheus Operator has been increased
to 1 GiB
-
Logging
- [CHANGE] Fluent Bit has been updated to version 1.7.9
- [FEATURE] Fluent Bit disk buffering is now enabled
- [FIX] Fluent Bit pods were not restarted properly during an
upgrade-in-place - [FIX] OpenShift: Upgrade-in-place now functions properly
- [FEATURE] OpenShift: A custom route hostname can be set with
OPENSHIFT_ROUTE_HOST_KIBANA
andOPENSHIFT_ROUTE_HOST_ELASTICSEARCH
1.0.8 - June 14, 2021
Version 1.0.8 (14JUN21)
-
Monitoring
- [EXPERIMENTAL] OpenShift automation
- Deployment to OpenShift clusters is now supported experimentally
- OpenShift support is still under development. Usage and features may
change until the feature set is finalized. - Documentation: Deploying Monitoring on OpenShift
- [FEATURE] The new
NGINX_DASH
environment variable now controls whether
the NGINX dashboard gets deployed when usingdeploy_monitoring_*.sh
or
deploy_dashboards.sh
.
- [EXPERIMENTAL] OpenShift automation
-
Logging
- [EXPERIMENTAL] OpenShift automation
- Deployment to OpenShift clusters is now supported experimentally
- OpenShift support is still under development. Usage and features may
change until the feature set is finalized. - Documentation: Deploying Log Monitoring on OpenShift
- [FEATURE] Container runtimes other than Docker are now supported.
- [EXPERIMENTAL] OpenShift automation
1.0.7 - May 17, 2021
Version 1.0.7 (17MAY21)
-
Overall
- Research was completed that will enable OpenShift support in a future release
-
Monitoring
- [CHANGE] Severtal component versions have been updated
- Grafana: 7.4.1 -> 7.5.4
- Prometheus: 2.24.1 -> 2.26.0
- Prometheus Operator: 0.45.0 -> 0.47.0
- Prometheus Operator Helm Chart: 13.7.2 -> 15.0.0
- kube-state-metrics: 1.9.7 -> 1.9.8
- [FIX] Upgrade-in-place of the Prometheus Pushgateway fails
- [FIX] CAS dashboard: Uptime widget format changed
- [FIX] CAS dashboard: Dashboard errors with some CAS configurations
- Instructions are now available
for manual cleanup if the monitoring namespace is deleted instead of running
the remove_* scripts
- [CHANGE] Severtal component versions have been updated
1.0.6 - April 19, 2021
Version 1.0.6 (19APR21)
-
Overall
- [FEATURE] Custom names for the NGINX controller service are now supported
via theNGINX_SVCNAME
environment variable (oruser.env
setting). - [CHANGE] Several updates to documentation have been made to improve clarity
and organize the content in a more useful way.
- [FEATURE] Custom names for the NGINX controller service are now supported
-
Monitoring
- [FEATURE] There is a new sample that demonstrates
how to enable Google Cloud's Operation Suite to collect metrics a Prometheus
instance that is scraping metrics from SAS Viya components - [CHANGE] The Amazon CloudWatch sample has been
updated to include many more metrics and mappings. Almost all metrics exposed
by SAS Viya and third party components are now mapped properly to sets of
dimensions. A new reference documents
the metrics by dimention, by source, and by metric name.
- [FEATURE] There is a new sample that demonstrates
-
Logging
- [FIX] Missing Kubernetes metadata on log messages from some pods (inc. CAS
server pod) has been fixed. Prior to fix, the kube.namespace field was set
tomissing_ns
and all otherkube.*
fields were not present.
- [FIX] Missing Kubernetes metadata on log messages from some pods (inc. CAS
1.0.5 - March 15, 2021
-
Overall
- There is a new document discussing support of various
Cloud providers
- There is a new document discussing support of various
-
Monitoring
- [FEATURE] The
monitoring/bin/deploy_dashboards.sh
script now accepts a
file or directory argument to deploy user-provided dashboards - [FEATURE] A new
$USER_DIR/monitoring/dashboards
directory is now
supported to supply user-provided dashboards at deployment time - [FEATURE] The new CloudWatch sample provides
instructions on configuring the CloudWatch agent to scrape metrics
from SAS Viya components - [FEATURE] The browser-accessible URL for Grafana is now included in
the output ofmonitoring/bin/deploy_monitoring_cluster.sh
(including
if ingress is configured) - [CHANGE] Several component versions have been upgraded
- Prometheus:
v2.23.0 -> v2.24.0 - Grafana:
7.3.6 -> 7.4.1 - Prometheus Operator:
0.44.1 -> 0.45.0 - kube-prometheus-stack:
12.8.0 -> 13.7.2
- Prometheus:
- [CHANGE] The following optional Grafana plugins are no longer installed by default:
- grafana-piechart-panel
- grafana-clock-panel
- camptocamp-prometheus-alertmanager-datasource
- flant-statusmap-panel
- btplc-status-dot-panel
- [CHANGE] cert-manager resources
now use 'v1' to align with their use in SAS Viya 4.x
- [FEATURE] The
-
Logging
- [FEATURE] The browser-accessible URL for Kibana included in the output
oflogging/bin/deploy_logging_open.sh
now takes into account ingress
configuration - [EXPERIMENTAL] A new experimental script
logging/bin/getlogs.sh
allows exporting logs to CSV formatDocumentation
- [FIX] The
logging/bin/change_internal_password.sh
script no longer
outputs passwords as debug messages
- [FEATURE] The browser-accessible URL for Kibana included in the output
1.0.4 - February 15, 2021
-
Overall
- Improved documentation for overall deployment process
- Improved documentation related to use of TLS
- Removed references to TLS in ingress sample (samples/ingress); TLS enabled ingress shown in TLS sample (samples/tls)
-
Monitoring
- [FIX] ENABLE_TLS should set proper port and targetport for v4m-prometheus service
- [FIX] Remove memory limit on kube-state-metrics
- [FIX] Kubernetes Cluster Dashboard disk usage not working on EKS
-
Logging
- Moved Helm chart from deprecated
stable/fluent-bit
tofluent/fluent-bit
- Fluent Bit version upgraded from 1.5.4 to 1.6.10
- Moved Helm chart from deprecated
1.0.3 - January 15, 2021
-
Overall
- Significantly improved documentation for deployment customization
KEEP_TMP_DIR
option added to keep the temporary working directory
around for troublshooting purposes- There is now an early check for
kubectl
cluster admin capabilities
-
Monitoring
- Component versions upgraded
- Helm Chart: 11.1.3->12.8.0
- Prometheus Operator: 0.43.2->0.44.1
- Prometheus: v2.22.2-> v2.23.0
- Grafana: 7.3.1->7.3.6
- The application filter on the SAS Java Services dashboard is now sorted
- The Perf/Node Utilization dashboard now uses node names instead of IP
addresses to identify nodes
- Component versions upgraded
-
Logging
- Moved Helm chart from deprecated
stable/elasticsearch-exporter
to
prometheus-community/elasticsearch-exporter
- Improved handling of log message fragment created due to excessively long
log messages (>16KB) - FIX: Eliminated hard-coded namespace in change_internal_password.sh script
- Moved Helm chart from deprecated