Metrics
Overview
Kloudfuse Platform supports high cardinality metrics from hundreds of sources including infrastructure (node cpu, memory, disk, etc), Kubernetes, and services like Kafka, Redis, MongoDB, Postgres, MySQL, etc. Kloudfuse comes with native integration with popular cloud platforms like AWS, GCP and Azure. It can ingest high cardinality metrics from applications without breaking the bank or requiring any special tiers or pricing.
Kloudfuse platform provides its native UI as well as integration with Grafana for exploring your metrics. The Kloudfuse stack ships with an inbuilt Grafana server that can be used to explore metrics using the PromQL interface or it can be configured as a data source in an existing external Grafana installation.
For users who prefer a simpler interface Kloudfuse stack ships with a native UI that lets folks explore the metrics and provides many functions to model modern dev-ops patterns. The Metrics UI can be accessed by clicking the “Metrics” tab in the UI header. Metrics UI provides three different ways to explore the data. Metrics Explorer lets the user do ad-hoc analysis of any metric by charting the time-series data for a selected range of time and dimensions. Metrics Summary provides a concise view of all metrics ingested by the system along with all the tags and their possible values within the time range selected. Metrics Dashboard lets the user create
Metrics Explorer
Metrics Explorer can be used to do ad-hoc analysis of various metrics ingested. The form builder allows users to select a metric name and optionally filter the metric using the “From” field. Multiple filters can be added using the field and they are ‘AND’d together. The next drop-down allows to aggregate the metrics across other space dimensions like host, cluster, or service. The default is to calculate the average across all the dimensions. If aggregation is not desired, then it can be removed by clicking the 'X' button.
Kloudfuse stack allows various other functions to be applied to the selected time series. They can be accessed by clicking on the button with the Sigma icon.
To write more complex queries click the “Add Query” button to chart another query and then combine multiple queries using the “Add Formula” button.
Kloudfuse UI allows users to view the equivalent PromQL queries. To switch to the PromQL command mode click the “< >” button next to the query form builder.
Kloudfuse supports various advance functions/algorithms for exploration of your metric data. These algorithms are also available for alerting when applicable in each stream as can be found here.
Metrics Summary
The metrics summary view allows the user to get a glance at all the tags (keys/values) reported for the metrics. The metrics are listed in a list view and can be searched using the provided search box. The Metrics Details subpage shows various metadata associated with a metric including where all the metric is reported from and its associated cardinality. The metric can be charted by selecting any of the labels that would be used as a filter.
Metrics Dashboard
Grafana Support
The Kloudfuse platform provides a fully compliant PromQL interface for exploring its metrics. This allows users to use standard open-source tools like Grafana to explore the metrics. Kloudfuse stack ships with an inbuilt Grafana server as well that can be accessed by clicking on the “Grafana” tab on the top bar. Alternatively, the user can configure an external Grafana by adding the kloudfuse stack as a Prometheus data source. Please refer to Grafana documentation for further details. Users can use the Grafana service to create custom dashboards and do ad-hoc metric exploration. Users can also import their existing Grafana dashboards to this instance. Please ensure that KfuseDatasource’s scrape interval parameter matches the actual scrape interval of the metrics data collection. Without this, queries can result in empty results.
Metric Source Integration
Kloudfuse supports ingesting metrics from various sources including cloud services, Kubernetes, platform services like Kafka, MySQL, and applications. Users can use a variety of ways to send the data to the kloudfuse stack. Here is a list of how to configure various sources.