add duration metric, tolerate missing tracepoints, ship el8 build

- Bubbles, treemap, icicle, and the live flamegraph 'b' cycle now include
  syscall duration (sum) as a third metric alongside events and bytes.
  Statsengine snapshots expose TotalLatencyNs to support this.
- AttachAll takes an optional warn callback. Production passes one so older
  kernels that lack newer tracepoints log a warning and keep going instead
  of aborting startup.
- Dockerfile.el8 + scripts/build-with-docker-el8.sh + mage buildDockerEl8
  produce ior.el8, a static binary built against Rocky Linux 8 glibc for
  RHEL/Rocky/Alma 8 hosts.
- README.md documents installing mage and the new el8 target.

4 files changed, 39 insertions, 33 deletions

diff --git a/internal/statsengine/filerank.go b/internal/statsengine/filerank.go
index d24ab93..ef43c6e 100644
--- a/internal/statsengine/filerank.go
+++ b/internal/statsengine/filerank.go
@@ -191,12 +191,13 @@ func (s fileSnapshotInput) toSnapshot() FileSnapshot {
 	}
 
 	return FileSnapshot{
-		Path:         s.path,
-		Accesses:     s.accesses,
-		BytesRead:    s.bytesRead,
-		BytesWritten: s.bytesWritten,
-		AvgLatencyNs: avg,
-		MaxLatencyNs: s.maxLatency,
+		Path:           s.path,
+		Accesses:       s.accesses,
+		BytesRead:      s.bytesRead,
+		BytesWritten:   s.bytesWritten,
+		AvgLatencyNs:   avg,
+		MaxLatencyNs:   s.maxLatency,
+		TotalLatencyNs: s.totalLatency,
 	}
 }
 
diff --git a/internal/statsengine/process.go b/internal/statsengine/process.go
index 3bfd019..b7eb6e7 100644
--- a/internal/statsengine/process.go
+++ b/internal/statsengine/process.go
@@ -174,11 +174,12 @@ func (s processSnapshotInput) toSnapshot(rateDiv float64) ProcessSnapshot {
 	}
 
 	return ProcessSnapshot{
-		PID:          s.pid,
-		Comm:         s.comm,
-		Syscalls:     s.count,
-		RatePerSec:   safeRate(s.count, rateDiv),
-		Bytes:        s.totalBytes,
-		AvgLatencyNs: avg,
+		PID:            s.pid,
+		Comm:           s.comm,
+		Syscalls:       s.count,
+		RatePerSec:     safeRate(s.count, rateDiv),
+		Bytes:          s.totalBytes,
+		AvgLatencyNs:   avg,
+		TotalLatencyNs: s.totalLatency,
 	}
 }
diff --git a/internal/statsengine/snapshot.go b/internal/statsengine/snapshot.go
index f2b617b..7a95ab8 100644
--- a/internal/statsengine/snapshot.go
+++ b/internal/statsengine/snapshot.go
@@ -68,12 +68,13 @@ type SyscallSnapshot struct {
 	Errors     uint64
 	Bytes      uint64
 
-	LatencyMinNs  uint64
-	LatencyMaxNs  uint64
-	LatencyMeanNs float64
-	LatencyP50Ns  uint64
-	LatencyP95Ns  uint64
-	LatencyP99Ns  uint64
+	LatencyMinNs   uint64
+	LatencyMaxNs   uint64
+	LatencyMeanNs  float64
+	TotalLatencyNs uint64
+	LatencyP50Ns   uint64
+	LatencyP95Ns   uint64
+	LatencyP99Ns   uint64
 }
 
 // FileSnapshot is an aggregated per-file ranking entry.
@@ -84,8 +85,9 @@ type FileSnapshot struct {
 	BytesRead    uint64
 	BytesWritten uint64
 
-	AvgLatencyNs float64
-	MaxLatencyNs uint64
+	AvgLatencyNs   float64
+	MaxLatencyNs   uint64
+	TotalLatencyNs uint64
 }
 
 // ProcessSnapshot is an aggregated per-process entry.
@@ -97,7 +99,8 @@ type ProcessSnapshot struct {
 	RatePerSec float64
 	Bytes      uint64
 
-	AvgLatencyNs float64
+	AvgLatencyNs   float64
+	TotalLatencyNs uint64
 }
 
 // HistogramBucketSnapshot is one bucket of a histogram snapshot.
diff --git a/internal/statsengine/syscall.go b/internal/statsengine/syscall.go
index 93931d1..2ef929a 100644
--- a/internal/statsengine/syscall.go
+++ b/internal/statsengine/syscall.go
@@ -195,18 +195,19 @@ func (s *syscallStats) ensurePercentiles() {
 
 func (s syscallSnapshotInput) toSnapshot(rateDiv float64) SyscallSnapshot {
 	return SyscallSnapshot{
-		TraceID:       s.traceID,
-		Name:          s.name,
-		Count:         s.count,
-		RatePerSec:    safeRate(s.count, rateDiv),
-		Errors:        s.errorCount,
-		Bytes:         s.totalBytes,
-		LatencyMinNs:  s.minLatency,
-		LatencyMaxNs:  s.maxLatency,
-		LatencyMeanNs: float64(s.totalLatency) / float64(maxU64(s.count, 1)),
-		LatencyP50Ns:  s.p50Latency,
-		LatencyP95Ns:  s.p95Latency,
-		LatencyP99Ns:  s.p99Latency,
+		TraceID:        s.traceID,
+		Name:           s.name,
+		Count:          s.count,
+		RatePerSec:     safeRate(s.count, rateDiv),
+		Errors:         s.errorCount,
+		Bytes:          s.totalBytes,
+		LatencyMinNs:   s.minLatency,
+		LatencyMaxNs:   s.maxLatency,
+		LatencyMeanNs:  float64(s.totalLatency) / float64(maxU64(s.count, 1)),
+		TotalLatencyNs: s.totalLatency,
+		LatencyP50Ns:   s.p50Latency,
+		LatencyP95Ns:   s.p95Latency,
+		LatencyP99Ns:   s.p99Latency,
 	}
 }