search

louislam / uptime-kuma

A fancy self-hosted monitoring tool
https://uptime.kuma.pet
MIT License
55.89k stars 5.03k forks source link

Down > Up Retries Option #3031

Open 1A3Dev opened 1 year ago

1A3Dev commented 1 year ago

⚠️ Please verify that this feature request has NOT been suggested before.

🏷️ Feature Request Type

Other

🔖 Feature description

If something being monitored is having intermittent issues it causes inaccurate statuses due to the fact it instantly gets marked as Up if it's detected as online so I think it could make sense to have another "Retries" option however instead of checking Up > Down this would check Down > Up.

For example if a website is marked as Down it would then require X amount of heartbeats to return as online before actually marking it as Up.

✔️ Solution

Rename the current "Retries" option to "Down Retries" and add a new option for "Up Retries" which does the same as the existing one however checks Down > Up instead of Up > Down.

❓ Alternatives

No response

📝 Additional Context

No response

CommanderStorm commented 1 year ago

I would argue that this extra bookkeeping is mostly unnecessary. In my experience, services mostly don't hover in the up-down state. => I would expect, you mostly want to get that notification that things are back okay as soon as possible after addressing the issue.

What is the exact use case you had a problem with Retries?

1A3Dev commented 1 year ago

What is the exact use case you had a problem with Retries?

I had a web server that was intermittently switching between HTTP 200 and 503 which was spamming notifications even with retries set to 2 which would be less spammy if this existed.

CommanderStorm commented 1 year ago

Okay, but that is something that should not happen => something you WANT to be notified for, right?

1A3Dev commented 1 year ago

Okay, but that is something that should not happen => something you WANT to be notified for, right?

No, the issue was caused by the hosting company where the web server was which I was unable to fix myself. You are correct that I'd want to know it's an intermittent issue but after I discovered that, I wanted a notification when it was fully back up which would be easier with this.

I do agree that it's not a necessary change however it would be a nice quality of life improvement.

I would say that maintenance mode would work to kinda fix the issue of notification spam. However with maintenance mode there's no way to see the current status of the service whilst the maintenance is active which means that isn't a valid option. If the response time graph via the dashboard showed the actual status rather than "maintenance" that would be good alternative solution.