8321ca25d7
Switch all stat panels in the hardware and hardware compression dashboards from lastNotNull to last reduceFunction. When a host is unreachable or node-proxy stops, metric silence is ambiguous — it could mean host down, agent crash, or scrape failure. lastNotNull freezes the last known value (e.g. 4 NVMe drives), causing inconsistency with the global Overview which reflects live count reductions. Use last so panels show N/A when no current data arrives. Fixes: https://tracker.ceph.com/issues/78360 Signed-off-by: Afreen Misbah <afreen@ibm.com> |
||
|---|---|---|
| .. | ||
| dashboards | ||
| dashboards_out | ||
| tests_alerts | ||
| tests_dashboards | ||
| .gitignore | ||
| .pylintrc | ||
| alerts.jsonnet | ||
| alerts.libsonnet | ||
| CMakeLists.txt | ||
| config.libsonnet | ||
| dashboards.jsonnet | ||
| dashboards.libsonnet | ||
| jsonnet-bundler-build.sh | ||
| jsonnetfile.json | ||
| jsonnetfile.lock.json | ||
| lint-jsonnet.sh | ||
| Makefile | ||
| mixin.libsonnet | ||
| prometheus_alerts.libsonnet | ||
| prometheus_alerts.yml | ||
| README.md | ||
| requirements-alerts.txt | ||
| requirements-grafonnet.txt | ||
| requirements-lint.txt | ||
| test-jsonnet.sh | ||
| tox.ini | ||
Prometheus Monitoring Mixin for Ceph
A set of Grafana dashboards and Prometheus alerts for Ceph.
All the Grafana dashboards are already generated in the dashboards_out
directory and alerts in the prometheus_alerts.yml file.
You can use the Grafana dashboards and alerts with Jsonnet like any other prometheus mixin. You can find more resources about mixins in general on monitoring.mixins.dev.
Grafana dashboards for Ceph
In dashboards_out you can find a collection of
Grafana dashboards for Ceph Monitoring.
These dashboards are based on metrics collected from prometheus scraping the prometheus mgr plugin and the node_exporter (0.17.0).
Prometheus alerts
In prometheus_alerts.libsonnet you'll find a set of Prometheus
alert rules that should provide a decent set of default alerts for a
Ceph cluster. After building them with jsonnet put this file in place according to your Prometheus
configuration (wherever the rules configuration stanza points).
SNMP
Ceph provides a MIB (CEPH-PROMETHEUS-ALERT-MIB.txt) to support sending Prometheus alerts to an SNMP management platform. The translation from Prometheus alert to SNMP trap requires the Prometheus alert to contain an OID that maps to a definition within the MIB. When making changes to the Prometheus alert rules file, developers should include any necessary changes to the MIB.
Multi-cluster support
Ceph-mixin supports dashboards and alerts across multiple clusters.
To enable this feature you need to configure the following in config.libsonnnet:
showMultiCluster: true,
clusterLabel: '<your cluster label>',
Building from Jsonnet
Method 1: System Packages (Recommended for most users)
sudo dnf install jsonnet jsonnet-bundler # RHEL/Fedora
sudo apt-get install jsonnet jsonnet-bundler # Ubuntu/Debian
# Install dependencies
jb install
# Generate all dashboards and alerts
make generate
# Run tests
make test
Method 2: Local Build (For CI/CD or no root access)
./jsonnet-bundler-build.sh
# Install dependencies using local jb
./jb install
# Generate all dashboards and alerts
make generate
# Run tests
make test
Ceph Mixin Development Commands
make all # Format, generate, lint, and test
make fmt # Format Jsonnet files
make generate # Regenerate JSON dashboards and alerts from templates
make lint # Run linters
make test # Run all tests
make vendor # Install dependencies
make help # List all available commands
Supported Versions
- jsonnet: v0.18.0+
- jsonnet-builder: 0.4.0+
- grafonnet-lib: always uses latest
masterbranch - alertmanager: 0.16.2+
- prometheus: v2.33.4+
