Better alert naming


(Giacomo Sanchietti) #1

Some of you asked for better alert translations on mail and Dartagnan UI.

@edoardo_spadoni is doing the work and we tried to write down the translation list.
Do you have any suggestion?

This is the actual alarm list with suggested translation:

  • alert_link: Server communicaition/heartbeat failed
  • alert_backup: Backup failed, Backup backup-name failed
  • alert_load: High system load
  • alert_service: Service service-name stopped
  • alert_df_boot: Boot partion full
  • alert_df_root: Root partion full
  • alert_swap: Swap partion full
  • alert_md: RAID raid-name failed
  • alert_ping: Ping latency hostname high
  • alert_droprate: Ping droprate hostname high
  • alert_nut: UPS on battery
  • alert_wan: WAN wan-name down

What do you think? /cc @danb35 @m.traeumner


(Dan) #2

It’s certainly an improvement, at least for the root/boot partition. Though it’s misleading, as the warning fires well before the partition in question is truly full, it at least gives a much better indication of the problem than “boot partition failed.”


(Giacomo Sanchietti) #3

We are going to deploy the modification in next few hours.

@danb35 you can change the alert configuration using this commands:

db alerts set 1 df FailureMin XX Instance root PluginType percent_bytes  TypeInstance free
/sbin/e-smith/signal-event nethserver-alerts-save

Where XX it’s the minimum free disk percentage which triggers the alert (default is 15%).
I’m sorry, but this part is not still documented.


(Michael Träumner) #4

I think a link to the log where you can find more information about the error could be nice.


(Giacomo Sanchietti) #5

Most of the time, you don’t have a specific log where to check the problem (see for example the disk usage case).

Often you need check multiple things inside the system.
I think a better solution is a link to a documentation for the specific problem which a detailed description and list of resources on how to solve it.