Skip to content

Return diagnostics snapshot on first Stats and refactor diagnostics collection #253

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
mateeullahmalik wants to merge 1 commit into from
Closed

Return diagnostics snapshot on first Stats and refactor diagnostics collection #253

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
63 changes: 56 additions & 7 deletions p2p/p2p_stats.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -114,6 +114,14 @@ func (m *p2pStatsManager) Stats(ctx context.Context, p *p2p) (*StatsSnapshot, er
}

prev := m.getSnapshot()
if prev == nil {
next, _ := m.refreshDiagnosticsSync(ctx, p)
if next != nil {
prev = next
} else {
prev = m.getSnapshot()
}
}
snap := cloneSnapshot(prev)
snap.PeersCount = peersCount
// Store a separate struct instance in the cache to avoid aliasing with the returned snapshot,
Expand All @@ -128,6 +136,45 @@ func (m *p2pStatsManager) Stats(ctx context.Context, p *p2p) (*StatsSnapshot, er
return snap, nil
}

func (m *p2pStatsManager) refreshDiagnosticsSync(ctx context.Context, p *p2p) (*StatsSnapshot, error) {
if m == nil || p == nil {
return nil, nil
}
if !m.refreshInFlight.CompareAndSwap(false, true) {
return nil, nil
}

defer m.refreshInFlight.Store(false)
start := time.Now()
refreshCtx, cancel := context.WithTimeout(context.Background(), p2pStatsRefreshTimeout)
next, err := m.collectDiagnostics(refreshCtx, p, m.getSnapshot())
cancel()
dur := time.Since(start)

if next != nil {
m.setSnapshot(next)
m.markFresh()
}

if err != nil {
logtrace.Warn(ctx, "p2p stats diagnostics initial refresh failed", logtrace.Fields{
logtrace.FieldModule: "p2p",
"refresh": "diagnostics",
"ms": dur.Milliseconds(),
logtrace.FieldError: err.Error(),
})
}
if dur > p2pStatsSlowRefreshThreshold {
logtrace.Warn(ctx, "p2p stats diagnostics initial refresh slow", logtrace.Fields{
logtrace.FieldModule: "p2p",
"refresh": "diagnostics",
"ms": dur.Milliseconds(),
})
}

return next, err
}

func (m *p2pStatsManager) maybeRefreshDiagnostics(ctx context.Context, p *p2p) {
if m == nil || p == nil {
return
Expand All @@ -142,10 +189,15 @@ func (m *p2pStatsManager) maybeRefreshDiagnostics(ctx context.Context, p *p2p) {

start := time.Now()
refreshCtx, cancel := context.WithTimeout(context.Background(), p2pStatsRefreshTimeout)
err := m.refreshDiagnostics(refreshCtx, p)
next, err := m.collectDiagnostics(refreshCtx, p, m.getSnapshot())
cancel()
dur := time.Since(start)

if next != nil {
m.setSnapshot(next)
m.markFresh()
}

if err != nil {
logtrace.Warn(logCtx, "p2p stats diagnostics refresh failed", logtrace.Fields{
logtrace.FieldModule: "p2p",
Expand All @@ -164,12 +216,11 @@ func (m *p2pStatsManager) maybeRefreshDiagnostics(ctx context.Context, p *p2p) {
}()
}

func (m *p2pStatsManager) refreshDiagnostics(ctx context.Context, p *p2p) error {
func (m *p2pStatsManager) collectDiagnostics(ctx context.Context, p *p2p, prev *StatsSnapshot) (*StatsSnapshot, error) {
if err := ctx.Err(); err != nil {
return err
return nil, err
}

prev := m.getSnapshot()
next := cloneSnapshot(prev)

var refreshErr error
Expand Down Expand Up @@ -203,9 +254,7 @@ func (m *p2pStatsManager) refreshDiagnostics(ctx context.Context, p *p2p) error
}
}

m.setSnapshot(next)
m.markFresh()
return refreshErr
return next, refreshErr
}

func cloneSnapshot(in *StatsSnapshot) *StatsSnapshot {
Expand Down
Loading