1 FMEA Worksheet ¶
This registry is a Failure Mode and Effects Analysis (IEC 60812 style) of the product: each record is one failure mode, scored for how bad it is, how often it happens, and how likely it is to escape detection before it reaches the user.
1.1 Scales ¶
Each factor is written into its own section of the record on a ten-point scale:
| Value | Severity | Occurrence | Detection |
|---|---|---|---|
| 1 | No noticeable effect | Practically never | Caught immediately, always |
| 4 | Degraded performance | Occasional | Usually caught by tests |
| 7 | Loss of function | Frequent | Rarely caught before release |
| 10 | Safety or data loss | Persistent | No detection in place |
The Initial RPN is Severity times Occurrence times Detection, computed over the factors as first analysed. The Residual RPN uses the Residual Severity, Residual Occurrence, and Residual Detection sections, scored after the mitigation is in place; a record without residual scores has not been mitigated yet and its Residual cell stays blank.
An RPN of 50 or below is acceptable, 200 or above is unacceptable and blocks the release until mitigated; the band between is tolerable while a mitigation is in progress.
1.2 Lifecycle ¶
Records move through Identified, Analysed, Mitigating, Accepted, and Closed. Every record whose current status is not Closed counts as open in the registries summary.
| # | Title | Failure Mode | Severity | Occurrence | Detection | Mitigation | Affected Documents | Status | Initial RPN | Residual RPN |
|---|---|---|---|---|---|---|---|---|---|---|
| FMEA-001 | Silent Data Corruption | A record saved while the storage is under pressure is written incompletely, without any error being raised. |
9 |
4 |
6 |
Write through to the storage layer and acknowledge only on its confirmation; add a per-record checksum verified on every read, so a corrupted record is refused instead of shown. |
REQ-005 | Mitigating | 216 | 36 |
| FMEA-002 | UI Freeze Under Load | The user interface stops responding for tens of seconds when a large data set is opened. |
5 |
5 |
3 |
Move the parse off the interface thread and show a cancellable progress indicator; analysis of the change is still in progress, so no residual scores are claimed yet. |
Analysed | 75 | ||
| FMEA-003 | Wrong Units in Report | A generated report renders a value in different units than its column caption states. |
8 |
3 |
4 |
Attach the unit to the value in the data model so the caption and the number can never come from different places; add a report regression test comparing every column against a golden file. |
REQ-006 | Mitigating | 96 | 16 |
| FMEA-004 | Failed Update Without Rollback | An interrupted software update leaves the installation in a state that neither starts nor accepts a new update. |
10 |
2 |
8 |
Not yet defined; the analysis of an A/B installation scheme with automatic fallback is scheduled for the next iteration. |
Identified | 160 | ||
| FMEA-005 | Log Overflow Fills Disk | Diagnostic logging grows without bound and eventually fills the disk the application runs on. |
4 |
4 |
2 |
A total cap with rotation was configured and shipped; the oldest files are dropped first and a warning is logged when the cap is reached. |
Closed | 32 | 8 |