DC/OS Monitoring v.1.3.0
- alertmanager: Add smtp auth secret support.
- alertmanager: Add a package option for git ref name.
- Add an option to configure prometheus remote write and remote read.
- Updated SDK to v0.57.3.
pod replaceoperation is required if there are existing deployments of older versions on DC/OS 2.0. See the SDK release notes for details at https://github.com/mesosphere/dcos-commons/releases/tag/0.57.3.
- Added option to configure the token refresh interval for Prometheus agent discovery task.
- Changed default token refresh interval for Prometheus agent discovery task to 1 hour.
- Upgraded version of Go to 1.13.
- Fixed typo in “dcos monitoring describe” CLI subcommand.
- Support Prometheus server placement constraints.
- Support scheduler placement constraints.
- Add new CLI subcommands for Prometheus (“dcos monitoring prometheus”).
- Add workaround for a bug in DC/OS versions 1.13.0-1.13.4 that caused permission errors in Nginx when running dcos-monitoring as root.
- Update SDK to v0.57.1.
- Fix the “More Detail” column in the Home dashboard in Grafana.
- Increase the default token refresh interval for Prometheus agent discovery task.
- Fix bug in agent discovery task causing leaked connections.
- Add a configuration option for the Grafana dashboard theme style (defaults to Light).
- Default DC/OS Grafana dashboards are loaded to a folder called DC/OS Defaults.
- Support setting pre-reserved roles for Prometheus.
- Decrease the number of resources in the default package configuration.
- Read in DC/OS CA certs for verification in the Prometheus agent discovery task (Enterprise only).
- Update SDK to v0.56.1.
- Remove configuration option for Grafana to use public node.
- Modify configuration option names for Secrets fields.
- Pushgateway is now optional and off by default.
- Installation is now blocked with helpful errors if extraneous configuration options are included.
- The DC/OS Monitoring CLI subcommand is now “monitoring” instead of “dcos-monitoring”.
- Upgrading from v0.5 is not possible due to Pushgateway defaulting to off.
- Add Marathon LB integration.
- Enable gzip compression for Grafana’s HTTP server.
- Add a link to the Grafana UI on the Services page.
- Set the Home dashboard in Grafana to the DC/OS Overview dashboard.
- Grafana dashboard titles are no longer overwritten by the file path of the dashboard json.
- Service name cannot contain characters that are not allowed in a DNS name (e.g., /) if using Marathon LB integration.
- Fix issues that occur when the service is installed in a folder.
- Fix a regression related to deploy key support.
- Workaround the UI bug for package install.
- Bug fixes (DSS profile support is missing).
- Allow accessing Alertmanager UI through Admin Router.
- Allow accessing Prometheus UI through Admin Router.
- Support loading Prometheus alert rules from a Git repository.
- Support reloading Alertmanager configurations after install.
- Support DC/OS Storage Service (DSS) volume profiles.
- Support for configuring Prometheus storage retention.
- Support for configuring Alertmanager resource footprint (i.e., CPU and memory).
- Installable from the Universe.
- Update SDK to 0.55.2.
- New logo.
- Upgrading from v0.3.0 is not supported at this time.
- Does not support loading Prometheus alert rules recursively from a directory.
- Allow accessing Grafana UI through Admin Router.
- Support configuring Grafana admin user credentials.
- Support Open DC/OS.
- Support Package Registry.
- Remove support for Prometheus Mesos Exporter, which is no longer used to collect Mesos metrics.
- Correct permissions required to run the package.
SSH authwhen fetching from Git repository.
- Add default Grafana dashboards which can be automatically loaded.
- Support Grafana server placement constraints.
- Support fetching from a branch in a Git repository.
- Grafana v5.3.4.
- Use local persistent volume for Grafana server.
- Use Docker image for Grafana server.
- Recursively clone Git sub-modules from a Git repository.
Bug fix release.
- Add missing release notes.
- Prometheus v2.2.1.
- Grafana v5.0.1.
- Alertmanager v0.15.2.
- Pushgateway v0.5.2.
- Automatically scrape dcos-metrics service endpoints in Prometheus.
- Automatically load Grafana dashboard configurations from a Git repository.
- Automatically load Alertmanager configurations from a Git repository.
- Support strict mode DC/OS cluster.
- Support launching Grafana server on public agents.
- No persistent storage for Grafana dashboard configurations.
- No external storage for Prometheus data.
- No backup for Prometheus data.