How can you know immediately if a NAS has an issue? After three nights of failed backup silence, I set up all the alarms on my phone

the time my mom's house tripped, the backup tasks on my old DS220j were stuck for three nights straight, and I didn't feel anything at all. On weekends, I logged in to check the log, and the task list had three red lines lying down. The earliest was early Tuesday morning.

I've written about this before: the about remote backups, I mentioned that the DS220j came to my house every day at 03:40 to load data, and when the air conditioner and microwave in my mom's old neighborhood were turned on, the switch would trip. It's not unusual for the circuit breaker to trip; what's strange is that after it jumped, no one knew for the first three days. On the task plan editing page, I clearly checked "Send only when the task is abnormally terminated," but later I realized—yes, I did check it, but that machine never had an email channel at all. It stirs up loneliness.

weekend, I paired three machines (my N100 Feiniu, my mom's DS220J, and my cousin's four-drive Synology slot) to alert each one. This post laid out the events, channels, and scripts—just copy them.

first think carefully about what alarm you want it to call before you start setting up the passage

many people set up alerts the opposite: first they study which app to use to push messages, then after setting up the alert, they find they don't know which events to subscribe to, and end up ringing eight times a day, all saying "system is running fine." First, list the events, rank them, then select the channel.

my rating is only in three tiers:

one more word about the hard drive. the precursor to a broken disk I wrote about abnormal noises, slow speed, and SMART count turning red—these warning signs are known first by the machine—the SMART data is stored in the drive's firmware. It knows, and if you don't look, it's as if no one knows. So hard drive alarms must be triggered by the machine itself, not just by waiting for you to remember and run for inspection.

one-sentence principle: you need to push your phone forward, glance at your email, and turn off what you don't need. all the other configurations revolve around this line.

Synology: One notification setup page, completing most of the

my mom's DS220j was the first to get one. In DSM, go to the Control Panel, find the "Notification Settings" tab and "Email." Service providers can directly choose QQ or choose custom SMTP to fill in :smtp.qq.com, port 465, SSL encryption.

here's a pitfall I fell into on my first try: you can't enter QQ passwords in the password bar. First, go to the QQ Mail web version, go to Settings and Accounts, enable POP3/SMTP services, and send a text message for verification. It will give you a 16-digit authorization code. In DSM, the password field is filled with this authorization code. To save myself, I entered my QQ password, clicked 'Send test email,' and the authentication failed. It took me ten minutes to check back and forth before I realized what was happening.

set up your email, and the same page also has a push service tab. Link your Synology account, install the DS finder app on your phone, and log in to the same account. System-level events like hard drive health, storage pool downgrade, UPS battery transfer, and DSM will automatically push it to your phone. It's much faster than email, and the lock screen pops up instantly. My cousin's four-disc Synology was remotely equipped with him. Same path, five minutes to finish.

finally returned to the source of the incident: the mission plan. On each task's editing page, there's a 'Send Email Notification' option. Check the box below, then 'Send only when the task is terminated abnormally'—after a normal run, it's not bothersome, only reported if it crashes. This combination works perfectly. But remember the lesson DS220j taught me: these checkboxes are just 'decide to report'; the channels in notification settings are the ones 'reporting with words.' If you don't have the right mouth, no matter how many decisions you make, you're still mute.

Feiniu: I couldn't find a systemized system layer, so I put the alert on the script side

my N100 runs on Feiniu, to be honest: in this version, I only dig for notifications in the settings, only a few switches, and don't have Synology's systematic notification center. Feiniu updates quickly; you might already have it by the time you see this article. Check it out yourself—I can't wait, the alert is done directly in Docker and the scripting layer.

two cron scripts. First, check the disk level, threshold 85%. This is the number the storage pool is full I've taken a loss on a 4TB drive. If it goes over, I'll send it to DingTalk. The second is to monitor hard drive health, run smartctl-H every Sunday morning, and if the output isn't PESSED, it triggers an alarm. How do you read the specific SMART numbers interpret that article is very detailed, and the script only recognizes the word PESSAID—rough but sufficient.

Docker here, the container upgrade post I mentioned I only had Watchtower monitor the two stateless containers, Jellyfin and Navidrome. Watchtower comes with built-in email notifications, and environment variables include WATCHTOWER_NOTIFICATION_EMAIL_FROM, WATCHTOWER_NOTIFICATION_EMAIL_TO, and SMTP server options. Every time the container is updated, it sends you an email. This is a 'just one glance' mode, using email instead of push notifications.

Two

Instantly Accessible Paths with a Phone: DingTalk Robot and Bark

emails are slow, QQ Mail normally takes a few seconds to a few minutes, and DS finder only handles Synology. The ones that truly arrived instantly across all platforms, I got two of them.

DingTalk robot. Create a group with only your own (start a group chat and add anyone and then kick them out, or create a group face-to-face). Group settings, intelligent group assistant, add bots, and choose custom settings. In the security settings, select "Custom Keywords" and enter the word "NAS"—note this trap: from now on, all incoming messages must include this word; if not, it will be rejected. After obtaining the webhook address, curl can be sent with just one line:

curl -s "https://oapi.dingtalk.com/robot/send?access_token=yourtoken" -H 'content-type: application/json' -d '{"msgtype":"text","text":{"content":" NAS disk water level alarm: 87% used "}}'

I tested it the first time after configuring it, and it returned {"errcode":310000,"errmsg":"keywords not in content"} At the time, I thought it was because the NAS couldn't get online. I checked the firewall for a long time and later realized it was because the message had forgotten to include the keyword "NAS." On Android phones, DingTalk is enough, and lock screen notifications pop up just as well.

if you're using an iPhone, add Bark. Install Bark in the App Store, open and copy your own Key, and at any time, a single line pushes: curl https://api.day.app/ your Key/NAS alert/disk level at 87% , lock screen instantly arrives, it uses Apple's native push notification, no need to open the app. It can even build its own server, which is an advanced feature—using the official version is totally enough. I gave Bark the backup cable, DingTalk tube the hard drive and water level, separated the two wires, and when they sound, you know what kind of thing it is.

my current alarm checklist, just copy it

all three machines are fully equipped, here's the current checklist:

water level script is only a few lines; just post it and adjust the threshold to use it:

#!/bin/bash
USE=$(df /vol01/1000 | awk 'NR==2{gsub("%",""); print $5}')
[ $USE -ge 85 ] & curl -s "https://oapi.dingtalk.com/robot/send?access_token=yourtoken" -H 'Content-Type: application/json' -d '{"msgtype":"text","text":{" content":"NAS disk water level alarm: used '"$USE"%"}}'

note that the script has the "NAS" prefix fixed, leaving space for the DingTalk keyword. Also, pick one day each month and temporarily set the threshold to 1% to run once a week. Check if your phone rings—the alert chain itself can be broken, SMTP authorization codes expire, DingTalk tokens reset, or CRON not running—all of which might make you think life is peaceful. Letting your phone ring is more reassuring than anything else.

don't turn the alarm into a wolf's turn

the most common mistake when it comes to this is subscribing too much. There are two types of posts you absolutely must not open:

First, login failure notifications. the SSH post I mentioned that during the half month I forwarded port 22, there were thousands of explosive logs. If every failed login had pushed my phone, I would have gone crazy. The correct way to fail login is to automatically block and regularly check reports, not to bomb in real time.

second, system update reminders. the conclusion of the about the system update is that important patches need to be upgraded, but the fact that "there is a new version" itself isn't worth interrupting you with. Just lie in your inbox and wait until you see it.

also has a double insurance tip: don't have only one channel for critical events. My mom's DS220j uses dual channel for email and DS finder. When I replied that QQ email had SMTP issues, the app push notifications were still alive, but the phone was supposed to ring. Don't put all your eggs in one basket, especially in the field of reporting the police.

finish with a clear order: first list the event list and classify the levels, then configure channels (Synology has a settings page, Feiniu uses scripts), and then check notifications for each task one by one, checking the link once a month. After installing this setup, the hard drive will fail, backups will fail, the switch will trip and power out—the phone will know before you do—those three nights of silent loss won't be endured again.