ReadMeFeed / Community knowledge
Real questions. Useful conversations. Find the people who know your stack.
Symptom Production services running Dropwizard 4.x generate excessive alert volume from health check state changes and high-cardinality metrics exported via Micrometer's Prometheus registry. The admin endpoint returns degraded status intermittently under load, while custom checks that call downstream dependencies cause cascading timeouts on the admin thread
Goal: Configure OpenShift Monitoring Alertmanager to forward only actionable alerts to the Red Hat Ansible Automation Platform webhook receiver, using inhibition rules to suppress low‑severity or duplicate noise before it triggers remediation jobs. Uncertainty: The documentation does not clarify whether Alertmanager inhibition rules are applied prior to send
On CentOS Stream systems deploying the documented Prometheus node_exporter and Alertmanager stack, operators must configure repeat_interval to balance re-notification against alert fatigue. node_exporter exports host metrics such as CPU, memory, and disk usage, while Alertmanager determines which alerts page on-call staff through grouping, inhibition, and si
Goal We want to maintain Alertmanager silence definitions across k3os upgrades so that noise‑reducing alerts remain effective after a new release is applied. Constraints The silences live in the k3s-alertmanager ConfigMap in kube-system , which is regenerated during an upgrade. There is no built‑in persistence for these custom rules, and the default Alertman
Goal The aim is to determine which alerting path—Grafana’s built‑in system or an external Prometheus Alertmanager—delivers dynamic recipient templating reliably while avoiding the use of production credentials. Constraints and Uncertainty Built‑in alerting supports a single‑step setup but currently lacks templated recipient support. Alertmanager integration
Goal Determine whether Alertmanager updates the annotations of an alert that Prometheus re‑emits with the same fingerprint but altered annotation values, especially after component restarts. Constraints and Uncertainty Alert identity is calculated from the alert name and label set, not from annotations. Alertmanager treats annotations as informational and do
Integration Boundary Prometheus HA replicas each expose firing and resolved alerts to a shared Alertmanager via the Prometheus Alerting API at /api/v1/alerts . Alertmanager deduplicates incoming alerts by comparing label sets, treating alerts with identical labels as a single group for notification purposes. Unresolved Behavior The deduplication logic is ver
prometheus
redhat
centos
grafana