[35421] Fixed the issue with StackLight sending a large number of duplicated alerts to Salesforce due to false-resolved notifications on the Prometheus side.
[35235] Fixed the expression and summary mismatch in the
SystemMemoryFullMajor
and SystemMemoryFullWarning
alerts.
[34980] Fixed the issue with parsing of multiline auditd logs in systemd inputs.
[34611] Fixed handling of Neutron metrics when using the LinuxBridge networking plugin. Renamed the Neutron Grafana dashboard OpenvSwitch panels to L2 and the Neutron OpenvSwitch Agent panel to Neutron L2 Agent.
[35339] Fixed the issue causing the following Salesforce-related services being unavailable due to inability to authenticate to Salesforce:
[35346] Fixed the issue with the CVP - StackLight tests
Jenkins pipeline job failing when running the test_nova_telegraf_metrics
test. The issue occurred due to the max_limit
setting, specified in
nova.conf
, restricting the number of retrieved servers from Nova. Now,
this setting is ignored.
[35343] Fixed the issue with the CVP - StackLight tests
Jenkins pipeline job failing with the TimeoutError: Incorrect image count in
metric error message when running the test_glance_metrics
test.
[35204] Parameterized the --web.cors.origin
flag to allow configuring
a CORS origin regex for Prometheus.
[35022] To avoid misinterpretation, the Telegraf metrics gathering alert was split into two separate alerts:
TelegrafGatherErrors
- a node-based Fluentd alertTelegrafRemoteGatherErrors
- a container-based Fluentd alert,
centralized through Docker SwarmFor details, see MCP Operations Guide: Telegraf.
[34821] Increased the prometheus-es-exporter
per-aggregate result
size to avoid situations when some logs are missing in Elasticsearch. Added
the ElasticsearchExporterNoDailyLogs
alert that raises when no new
published logs from a given node are detected by the Elasticsearch exporter.