Field notes / complete index

Things I had to learn the inconvenient way.

I am learning this as I go, not writing from the position of a professional network engineer or storage expert. Some notes describe failures and decisions in my own setup. Others unpack a useful technical question with an illustrative scenario; they are labelled as explainers and do not pretend to be firsthand incident reports.

The archive contains nine notes. I try to explain what a thing does, why someone would choose it, what a simpler alternative offers and how far the evidence supports the conclusion. The details that identify or operate the private network are not part of that lesson.

09Monitoring note

Operations

Why a green dashboard can sit beside a broken service

A monitor can accurately report that a page answers while the job behind that page is failing. This explainer shows what each kind of check can and cannot establish.

Short versionMonitor the user-visible job and its shared dependencies, not only whether a process answers a basic request.
08Security note

Networking

How to read a frightening UniFi security alert

A severe-looking notice does not, by itself, prove that someone got in. This guide explains the difference between a detection, a block and evidence of actual access without describing my private exposure rules.

Short versionA blocked alert is evidence that something matched a rule and was stopped; it still deserves review, but it is not proof of a successful intrusion.
07Explainer

Storage

Why deleting a large folder may not return the space

A folder can disappear from the normal view while a recycle bin, snapshot or older version still retains the data. This explains a common storage puzzle; the example below is illustrative, not a claim about a specific deletion in my home.

Short versionWhen free space does not return after deletion, check recycle bins, snapshots, version history and cleanup progress before deleting more things in a panic.
06Explainer

Backups

A completed backup is not the same thing as a recovery plan

A successful backup job is reassuring, but it cannot answer the question that matters after a failure: what can actually be restored, and in what order? This is a planning guide, not a report of a completed disaster-recovery test.

Short versionA backup becomes useful only when its contents, location and restoration steps are understood and checked.
05Diagnostic note

Media

The Plex server is not always the problem

The same server and file can behave perfectly on one playback device and badly on another. Useful diagnosis begins by naming the exact device, app and playback mode instead of treating Plex as one indivisible machine.

Short versionInspect the active session and run controlled comparisons before changing the server, replacing the file or accusing the network on principle.
04Recovery note

Storage

When a network storage problem made the whole server look frozen

The server still had power and several local parts were running. Too many important actions were waiting on the same unavailable storage connection, which made one fault look like a failure of the entire platform.

Short versionRestore the shared storage from the bottom of the dependency chain upward before restarting everything that happens to look broken.
03Media note

Media

Why most of my library is still 1080p

The main television has a 4K screen, but screen resolution is only one part of playback. The file, app, audio, subtitles, network and server all get a vote, and several of them are surprisingly opinionated.

Short versionI use dependable 1080p as the default and choose 4K for individual films after confirming that the complete playback setup can use the extra quality.
02Explainer

Networking

What should happen to an app when its VPN disconnects?

A VPN can protect the route an application uses, but a connected icon tells only part of the story. This is a guide to the failure behaviour worth checking, not a report of a completed test on my network.

Short versionIf an application genuinely requires a VPN, check whether it loses access when the VPN stops; do not assume the regular connection is unavailable.
01Failure report

Networking

The network upgrade I eventually turned back down

I added a faster network connection to the NAS. File copies improved immediately. Reliability then deteriorated so badly that the faster storage sometimes disappeared altogether and made several unrelated services look broken.

Short versionThe NAS is still using the slower connection. I am leaving performance on the table, but the table no longer vanishes during dinner.

What counts as a note

Enough context to be useful.

An incident report needs more than a symptom and a setting that happened to change it. It should explain the expected behaviour, the observed failure and which shared dependency could connect the symptoms.

An explainer asks a different question: when would you use this tool, what would you check and what alternative might be simpler? It should not borrow the voice of an incident report to make a hypothetical example sound more exciting.

What these are not

Universal instructions or professional advice.

The equipment, software versions and surrounding network affect every result. A fix that worked here is evidence from one environment, not a guarantee for every reader with the same product name.

The notes aim to improve the reader's questions and testing order. They should not encourage copying a private configuration line by line, even if publishing that configuration would make the page conveniently longer.

How the archive grows

Slowly, after something teaches a lesson.

New notes can come from an actual problem, a decision worth explaining or a common question I had to understand. There is no posting quota. If the equipment behaves for a month, the site may enjoy the same quiet period.

Existing notes may be expanded or corrected when later evidence changes the explanation. The update log records meaningful revisions so the archive can learn without pretending the first version arrived engraved in stone.