prom-dashboard

This tool can be uses to fetch prometheus metrics from a list of specified CI runs in PROW, and outputs the data in a csv.

Usage

Running it is simple, it takes a yaml file and a directory as input. The directory should be empty.

go run main.go -c <yaml config> -o <metadata and output dir>

The yaml supports the following customizations:

// testIDs lists the ID of the CI Prow jobs you want to pull data from
testIDs:
    - 42
    - 41
    - 40

// promMetrics lists the prometheus metrics that you want to gather for every test in testIDs
// +optional, defaults to: [
//      "etcd_disk_wal_fsync_duration_seconds_bucket",
//      "etcd_disk_backend_commit_duration_seconds_bucket",
//      "etcd_network_peer_round_trip_time_seconds_bucket",
//  ]
promMetrics:
    - etcd_disk_backend_commit_duration_seconds_bucket

// Step allows you to set the step for ranged queries
// +optional: default: "1m"
step: 5m

Output

The final output of a run will be written to output-dir/results.csv. This csv file has the following schema:

| TestID | Metric | Start Time | End Time | Step | Node | Time Series Data | | --- | --- | --- | --- | --- | --- | --- | --- |

The Time series data is in time differentials based on the step you provided. So the first cell is 0 steps from the start time, and the second is +step. The data ends at the specified end time.