Strange uptime reporting

[seamushroon54]

I’m currently running 6 instances on a computer, the spec is as below:
R5-3600/16G, VDSL2 100M/40M /w static ip, dns manually set to 8.8.8.8on both computer, router, and docker, 1300UHP dedicated to this usage(pppoe), HDD is 4TB each(brand new, similar spec)
All nodes are setup at about the same time, but today I got something strange
which is some of the node’s up time for all satalites are ~99%, but some of which got ~50% only
I always use a up time bot to monitor each node’s port, and it showed ~99.8% for each nodes on this month, which some how contradicts to the up time report mentioned above.

For the convinience of investigation, here is the node it:
1SuiTT3d1iwmbsx7EKatcZpE6FvA8StXiiMNVvwaRiXujqqLRi
1zQv6YaSXJ1xmtyh4DE4Ww62avQMvTQiKABbhwVBf2Ra1BR8gp
12AzzVJvYcypka5yqu7hox5t46cF9A9MB59G7xPUg6frrBd6pFQ
12th7LWFqFfu4RN5jMwTdpd7sB8eZMXmWMNd8xd4SeaJ7ehF6cB
1EPzcZP7BHGGnDHtc12WVp4kvfPua43r3MyxyzrU1wPavAMQ33
1U9HiNgpau5yhKs9nH6qbBAYiEPkMw6SmGvNHY17on1o6rZsRw

what would be the potential cause? and how should I improve it? thanks.

[Alexey]

Hello @seamushroon54,
Welcome to the forum!

The “Uptime checks” on dashboard is a lifetime based, not spot time, and it doesn’t used, it’s under construction:

Blueprint: tracking downtime with audits

Hey everyone, we have a new design here for tracking downtime with audits We had a previous discussion on a related document here Blueprint: Downtime Disqualification , but we decided to reconsider using audits to measure storage node downtime. Let us know what you think

Regarding not 100% uptime on uptimerobot - more like your router is a reason.

[BrightSilence]

The current uptime system doesn’t impact your node like Alexey mentioned. Regardless I can think of 2 things that may cause this low percentage. Maybe these nodes briefly made contact with the satellite before you finally set them up. As soon as the satellite knows about the node, it will start tracking its uptime. So if you started it once and ran into some issue and only got around to fixing that weeks later. You are starting with a lot of down time.

The other possible issue Alexey already hinted to as well. Your router may have some issue which causes it to be dropping some connections. This is mostly invisible with other internet use as most stuff is retried automatically. I had this issue with my router. It also showed a lot of errors in the log. The only way I managed to fix it was to factory reset the router. I still don’t know the actual cause. But my percentages dropped below 90% because of this.

[seamushroon54]

@Alexey
Does it mean that Audit Checks in the bashboard is the only meaningful figure?

@BrightSilence
I had checked the log of uptime robot, the only disconnect time in this month is caused by a maintenance(router update).

Thanks for your help.

[BrightSilence]

Uptime robot didn’t catch the router issues for me either. It did show quite an erratic graph with long response times from time to time.

[seamushroon54]

Thanks for your info, I’ll do a detailed investigation tomorrow.

[Alexey]

@seamushroon54

Does it mean that Audit Checks in the bashboard is the only meaningful figure?

No, it is a real audit score, you can read here how it’s calculated:

• audit-v2.md
• audit-containment.md
• audit-suspend.md
• disqualification.md

[twl]

@BrightSilence

Uptime robot didn’t catch the router issues for me either. It did show quite an erratic graph with long response times from time to time.

How long would “long” be in that case? I see above 300ms on all my nodes iirc, all on VDSL lines and with wired LAN connections

[BrightSilence]

In my case seconds. I think 300ms is well within the normal range. Strangely enough one of my nodes seems to be showing weird spikes right now. Not sure what the problem is, my other nodes look fine. But if you see spikes like this, but much more often. That’s what I was talking about.

I guess I’ll have to make sure there’s nothing iffy going on with this node.