feat(rtt): 'machine rtt' command to show round-trip time between macines usign using gossip data (#226)

* Add corrosion admin client function to get RTT to other machines in the cluster

* Add `uc machine rtt` showing all pair-wise RTT stats from corrosion

* Add long description to machine rtt command

* Include machine peer RTTs in InspectMachine instead of adding new gRPC API for it

* Add nil check on InspectMachine's Rtt field to protect from a potential edge case

* Generate docs for machine rtt command

* Handle m.Message being nil when a node is down or unavailble

* Update cli-docs

* Use tui Table instead of tabwriter

* Change reported RTT to the median rather than mean. Still calculate include stddev, as it might be a useful indication of network jitter.

* Show RTT to peers in `wg show` output

* Update cli docs
This commit is contained in:
Justin Bradford
2026-04-20 08:01:20 +10:00
committed by GitHub
parent 1c0d48cb46
commit 8d023f5c53
10 changed files with 564 additions and 156 deletions
+105
View File
@@ -6,8 +6,10 @@ import (
"errors"
"fmt"
"io"
"math"
"net"
"net/netip"
"sort"
"time"
)
@@ -256,3 +258,106 @@ func (c *AdminClient) ClusterMembershipStates(latest bool) ([]ClusterMembershipS
}
return states, parseErr
}
type MemberRTTStats struct {
Addr netip.AddrPort
Median float64
StdDev float64
}
// ClusterMemberRTTs returns the median and standard deviation of round-trip times to each cluster member.
func (c *AdminClient) ClusterMemberRTTs() ([]MemberRTTStats, error) {
respCh, err := c.SendCommand([]byte("{\"Cluster\":\"Members\"}"))
if err != nil {
return nil, err
}
var stats []MemberRTTStats
var parseErr error
for r := range respCh {
if r.Err != nil {
return nil, r.Err
}
addr, rtts, err := parseClusterMemberRTT(r.JSON)
if err != nil {
parseErr = errors.Join(parseErr, err)
continue
}
if len(rtts) == 0 {
continue
}
sort.Float64s(rtts)
n := len(rtts)
var median float64
if n%2 == 0 {
median = (rtts[n/2-1] + rtts[n/2]) / 2
} else {
median = rtts[n/2]
}
var sum float64
for _, rtt := range rtts {
sum += rtt
}
avg := sum / float64(n)
var varianceSum float64
for _, rtt := range rtts {
diff := rtt - avg
varianceSum += diff * diff
}
stdDev := math.Sqrt(varianceSum / float64(n))
stats = append(stats, MemberRTTStats{
Addr: addr,
Median: median,
StdDev: stdDev,
})
}
return stats, parseErr
}
func parseClusterMemberRTT(json map[string]any) (netip.AddrPort, []float64, error) {
var addr netip.AddrPort
var rtts []float64
var err error
// Parse state to get Addr
stateObj, ok := json["state"].(map[string]any)
if !ok {
return addr, nil, fmt.Errorf("missing or invalid 'state' field")
}
if addrStr, ok := stateObj["addr"].(string); ok {
addr, err = netip.ParseAddrPort(addrStr)
if err != nil {
return addr, nil, fmt.Errorf("parse 'addr' field: %w", err)
}
} else {
return addr, nil, fmt.Errorf("missing or invalid 'addr' field in 'state'")
}
// Parse RTTs
if rttsVal, ok := json["rtts"]; ok {
if rttsSlice, ok := rttsVal.([]any); ok {
for _, v := range rttsSlice {
if f, ok := v.(float64); ok {
rtts = append(rtts, f)
} else {
return addr, nil, fmt.Errorf("invalid rtt value type: %T", v)
}
}
} else {
return addr, nil, fmt.Errorf("invalid 'rtts' field type: %T", rttsVal)
}
} else {
return addr, nil, fmt.Errorf("missing 'rtts' field")
}
return addr, rtts, nil
}