WordPress Reliability & Recovery Engine
FaultPilot
Detect. Diagnose. Prevent. Recover.
Your WordPress site should not wait for a failure to tell you something is wrong. FaultPilot continuously observes site health, identifies failures and anomalies, provides evidence-based diagnostics, helps prevent recurring incidents, and gives you controlled recovery mechanisms when things go seriously wrong.
Simulated data — illustrative console, not live production output
The reliability layer
Not another uptime checker
Ordinary monitoring tells you that something broke — after visitors already noticed. FaultPilot is an intelligence layer that lives inside WordPress: it normalizes raw errors into fingerprinted signals, groups them into incidents, correlates them with the changes that caused them, and hands you a diagnosis with evidence instead of a log dump.
How FaultPilot thinks
From raw signal to verified recovery
Every piece of evidence moves through a ten-stage pipeline. Each stage is a real subsystem in the engine — not a marketing slide.
Observe
Collectors hook the error handler, shutdown, REST, HTTP API, cron and mail — zero front-end overhead.
Detect
Events are normalized, fingerprinted, buffered and batch-written to stay lightweight.
Correlate
Related errors merge into incidents; change events within the lookback window become suspects.
Diagnose
Weighted evidence produces a primary suspect with an explicit confidence tier.
Assess Risk
Six risk components with per-factor explanations — no opaque scores.
Prevent
Hourly trend sweeps catch rising error rates, cron backlogs and autoload growth early.
Alert
Dashboard notification center and branded HTML email — deduplicated, throttled, quiet hours.
Recover
Safe Mode, single-use recovery links, recovery points and verified treatments.
Verify
Post-treatment checks confirm the fix — health signals re-measured, not assumed.
Learn
Baselines, activity logs and reports turn every incident into institutional memory.
Core capabilities
An engineering product, not a plugin checklist
Each capability below is a real subsystem with its own data model, UI and honest limits. Explore them in the interactive demo.
Intelligent Error Engine
PHP fatals, warnings, notices and deprecations; database errors; REST failures; outbound HTTP failures; cron problems; mail failures; admin JS errors — normalized, deduplicated and sampled.
Root Cause Analysis
Weighted evidence — recency, component match, occurrence spikes — ranked into a primary suspect with a confidence tier: Strong correlation, Likely, Possible, or Insufficient evidence.
Incident Management
Fingerprinted events group into incidents with lifecycle states, occurrence counts, severity, evidence lists and per-incident recommendations.
Prevention Engine
Hourly sweeps watch error trends, cron backlog, autoloaded option growth, memory trajectory, REST and mail streaks, SSL expiry and checkout reliability — before they become incidents.
Risk Analysis
Six explainable components — reliability, updates, database, performance, compatibility, security — each factor shipped with the score. No black-box health numbers.
Smart Alerts
Dashboard notification center plus branded HTML email with deduplication, throttling, minimum severity thresholds and quiet hours.
Change Timeline
Plugin, theme and core updates; activations; config changes — each recorded with automatic pre-update recovery snapshots.
Plugin & Theme Intelligence
Component health, compatibility signals, outdated WooCommerce template overrides and per-plugin error attribution across every incident.
Database Doctor
Table sizes, autoloaded options, transients, orphaned meta — with explain-first workflows. FaultPilot never deletes table data automatically.
Performance & Cron Doctors
Memory headroom, slow cues, REST loopback health, duplicate cron events, stuck schedules and Action Scheduler backlog — measured, not guessed.
WooCommerce Intelligence
Checkout reliability, payment gateway errors, webhook failures, HPOS-aware order signals and background job health. Monitors and diagnoses — never touches orders.
Recovery Architecture
Safe Mode quarantine drop-in, single-use HMAC-signed recovery links, recovery points, rollback and post-recovery verification.
Incident experience
What an incident looks like
This is the actual information model FaultPilot produces — evidence, correlation and recommended actions. The buttons below the incident switch the evidence panel, exactly like the product's incident detail screen.
Evidence
- PHP stack trace — Uncaught TypeError: str_replace(): arg #2 must be of type array, string given in includes/class-emailer.php:102
- Affected component — attributed to plugin woocommerce-emailer v3.4.1 by file path
- Recent change — plugin updated 3.4.0 → 3.4.1 at 14:02:31, first failure 14:32:08
- Repeated occurrence — 37 hits across front-end and REST, same fingerprint, deduplicated
- Related failure — 12 Action Scheduler jobs for the same hook failed after the update
Root cause analysis
confidence Strong correlation
evidence change match: 1.00 · temporal: 29m · spike: +37/h
weighted component match 0.42, temporal 0.31, spike 0.19, repeat 0.08
recommendation rollback to 3.4.0 from recovery point, or patch arg #2
- The analyzer states its confidence and the factors behind it — it never claims certainty it does not have.
Change timeline
- 14:02:31 — woocommerce-emailer updated 3.4.0 → 3.4.1 (pre-update recovery point taken 14:01:47)
- 14:31:55 — WP Object Cache flushed by cache plugin
- 14:32:08 — first occurrence of FP-82A19C (front-end request)
- 14:33:02 — REST batch failures for the same fingerprint (merged)
Simulated incident — demonstrates the real information model
Recovery
When it breaks, get it back
wp-admin is often the casualty of a serious failure — which is why FaultPilot's recovery architecture works outside the admin: a Safe Mode drop-in, single-use signed recovery links and verified treatments.
Failure detected
A plugin or theme fatal takes the site down — including wp-admin itself.
FaultPilot identifies the incident
The failure is fingerprinted, grouped and attributed to a component with evidence — even if the admin UI never rendered.
Single-use recovery link
An HMAC-signed, expiring, rate-limited link (created earlier, stored only as a hash) arms Safe Mode without touching wp-admin.
Problematic component isolated
The drop-in quarantines the failing plugin by filtering the active plugin list — no remote commands, no secrets inside, deletable at any time.
Site restored
Optionally roll back to a pre-update recovery point — the exact plugin set captured before the change.
Verification, not hope
Post-treatment checks re-measure health signals and confirm the fix. The token is consumed, the action is audit-logged.
WooCommerce
Store reliability, diagnosed like infrastructure
When checkout breaks, every minute is revenue. FaultPilot activates WooCommerce intelligence automatically and watches the failure surfaces stores actually hit — without ever touching orders or financial data.
- Checkout reliability — measured request outcomes, deterioration alerts
- Action Scheduler — backlog depth, failed jobs, stuck hooks
- Payment gateways & webhooks — HTTP failures attributed with evidence
- Plugin conflicts — errors attributed to the exact component
- Template overrides — outdated WooCommerce overrides reported
- HPOS-aware order signals — respects High-Performance Order Storage
Simulated data
Security & privacy
Reliability software must not become the attack surface
FaultPilot treats itself as security-sensitive software — least privilege, audited recovery, minimized data.
Least privilege
A dedicated capability guards every read and write, with nonce validation on all state changes. Nothing else can touch FaultPilot data.
No backdoors
The Safe Mode drop-in contains no secrets and accepts no remote commands — it reads one option and filters the plugin list. Deleting it disables the feature.
Single-use recovery links
256-bit tokens, HMAC-hashed at rest, expiring, rate-limited and audit-logged. The link is the credential for exactly one quarantine action.
Documentation
Written like it will be read at 3 a.m.
Installation, configuration, incidents, recovery, safe mode, alerts and honest troubleshooting — reflecting what the product actually does.
Install & baseline
Requirements, installation, onboarding and the first full diagnosis.
Incidents & recovery
Incident lifecycle, evidence, Safe Mode, recovery points, verification.
Diagnostics suite
All 16 checks, what they measure, and what they honestly cannot observe.
Security & FAQ
Threat model, data minimization, troubleshooting and limitations.
Stop waiting for failures to introduce themselves
Install the Free Edition, run the first diagnosis, and see your site the way FaultPilot does.