deep profile + threshold gating + firmware stage + Burn super-stage
CI / Lint + build + test (push) Failing after 1m57s
Release / release (push) Has been cancelled

Ships all five phases of the deep-profile overhaul together. Runs now
carry a profile (quick/deep/soak); every profile walks the same
11-stage order — Inventory → Firmware → SpecValidate → SMART →
CPUStress → Storage → Network → Burn → GPU → PSU → Reporting —
with only per-stage durations and concurrency scaled.

Phase 1: profiles.ProfileRegistry loaded from vetting.yaml; runs.profile
column + CreateWithProfile; threshold table + evaluator seeded per-run
from the shared vetting.thresholds block; breach flips result at
/sensor + /result.

Phase 2: upgraded CPUStress (stress-ng --cpu-method=all --verify +
EDAC/MCE poll), Storage (fio --verify=md5 + SMART start/end delta),
Network (sustained iperf + /proc/net/dev deltas) with per-profile
knobs from Deps.

Phase 3: Burn super-stage with goroutine fan-out for CPU + memory +
fio + iperf, PSU rails sampled across the Burn window, SensorMux
(2 s flush, 500-sample cap) to absorb backpressure.

Phase 4: Firmware stage + firmware_snapshots table; probes dmidecode
(BIOS), ipmitool (BMC), ethtool -i (NIC), nvme (sysfs + id-ctrl),
lspci (HBA), /proc/cpuinfo (microcode). spec.DiffFirmware folds into
SpecValidate with pin-by-identifier and fan-out-across-component
matching; mismatches park the run in FailedHolding.

Phase 5: profile radio on the host start form, profile chip on the
run header, Firmware section in the HTML report, coverage artifact
uploaded from CI, agent/tests/fakes/ scaffold with Deps.LookPath
seam + stress_ng and dmidecode example fakes.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This commit is contained in:
2026-04-18 22:50:57 -04:00
parent fbb21cbafd
commit 23c689aa5b
60 changed files with 5911 additions and 527 deletions
+25
View File
@@ -102,6 +102,21 @@ templ HostActions(d HostPageData) {
<div class="host-actions-row">
if hostCanStart(d) {
<form method="post" action={ templ.SafeURL(fmt.Sprintf("/hosts/%d/start", d.Host.ID)) } class="inline host-start-form">
<fieldset class="host-profile-picker">
<legend>Profile</legend>
<label title="~10 min — post-repair sanity: all probes + gates, short budgets">
<input type="radio" name="profile" value="quick" checked/>
quick
</label>
<label title="~812 h — overnight soak: long CPU/RAM, full-disk fio verify, 30 min network">
<input type="radio" name="profile" value="deep"/>
deep
</label>
<label title="≥24 h — week-long burn-in; opt-in when you suspect intermittent faults">
<input type="radio" name="profile" value="soak"/>
soak
</label>
</fieldset>
<label class="host-nd-toggle">
<input type="checkbox" name="non_destructive" value="1"/>
Non-destructive (skip wipe-probe + disk writes)
@@ -258,6 +273,16 @@ func hostCanStartIfOnline(d HostPageData) bool {
return d.ActiveRun == nil
}
// profileChipValue normalizes a Run.Profile string for display on the
// run page chip. Older runs with an empty column predate Phase 1 — show
// them as "quick" (the prior implicit default).
func profileChipValue(p string) string {
if p == "" {
return "quick"
}
return p
}
// runDuration formats the elapsed time for a run using the same buckets
// as stageDuration. In-flight runs clock from StartedAt to now so the
// run-page header + runs-table row keep ticking on each SSE push.