Community Feedback on Minting Overview Tool

[sabrinasadik]

Dear friends,

A couple of months ago, we launched a GEP in which we proposed removing the minimal uptime requirement for farming until the right tools to monitor your nodes are in place.

As we specified on that GEP, it is very important to strive for 100% uptime to achieve a stable and trustworthy TF Grid. A node that is down too much is not usable by anyone so there is no reason to reward the farming.

The team is working on creating a minting overview tool on which farmers will be able to see the payouts of the previous month, as well as the uptime of the current minting period. To make sure the tool is efficient enough, we would like to ask for the community’s feedback for this tool.

What do you, as a farmer, find important to be able to see on this minting overview tool?

Keep in mind that ThreeFold is working on a decentralized Grid. This means that we are not implementing any centralized service for this.

Thank you for the input!
The ThreeFold Team

[LongStreetTech]

Let’s start with something basic that actually works before we get too feature rich. I couldn’t even get the slow a** explorer to load info for me to check if farms, and which, were online the other day. I know it has to poll a lot of info, but maybe regular polling that is cached so you don’t have to wait forever? And for it to actually work…. Then any kind of analysis/deeper dive into hardware available maybe? I can’t dream that big yet. Added some storage to a node that day and when I rebooted it had a different ID. So there went it’s first two weeks of running down the drain, so all for no uptime requirement until some proper tools to monitor your farm(s) are available. The threefold app has one of my farms and it says nodes are online that have been offline for weeks. So I can’t trust it to help keep an eye on things. Oh, and I am upgrading processors in that same system as soon as I have time, so I guess another new ID? Should have saved the power and left it off. When I’m not super tired, would love to be more constructive on this topic, as proper monitoring and analysis tools are imperative to help maintain maximum uptime and efficiency. Thanks for bringing it up! I still have server upgrades to do to maximize each one and right now have to depend on the explorer to maybe, hopefully, give me vague ideas on my current hardware so I know what it is and if it needs changing. I mean, cores to RAM is pretty easy, but we go a little deeper and it could fold into system performance metrics. Identify and benchmark hardware or something. Useful for both farmer and renter. And I say that not as the DDR4 brand new server gentleman the other day. Bc I’m running ddr3. But bc it would be useful for all.

Okay, I’m done for now. Lol. Thanks!

[cdcreek]

As aand farmer I would like to see everything I need to manage the farm from the tf app. Including alerts that would be sent to me should any nodes go offline But what I really like to know is how the movement is GROWING. We (farmers) need more information on grid sales. The recent news on the partnership with Open Nebula is good…very good it tells us there are customers signing up for the services.
The dubai real estate partnership was another positive. Sales and marketing effectiveness will be crucial to our success so please provide updates on a regular basis to keep everyones confidence high.

[joey]

it would be nice if you can see the status of your node in the app whether it is online or not

[scott]

This is on the roadmap for the TF Connect app. In the meantime, you can check your node’s status using the Telegram bot: Announcing the node status bot for Telegram

[RobertL]

Just a thought here; i fully understand the need for a stable network with 100% uptime or close. However, as i already expected reading one of the above posts; if for some reason (upgrade, replacement, relocation) a node is offline for let’s say 1%, if this means receiving no rewards at all, with current power prices I’d definitely leave it switched off for the remaining month, decreasing the stability of the Grid even more.
More on subject;

• which nodes are online (close to real time)
• type of node (presearch, certified etc.)
• growing rewards
• percentage/expected rewards at the end of the month based on current uptime
• maybe in a later stage the actual utilisation of the node would be good, especially when this is rewarded too

Hope this helps

[ParkerS]

your node id doesn’t change just from upgrades, your nodes id is stored on the first ssd it boots, it didn’t stay because you didn’t follow the guide in the faq on backing up your node id. If you haven’t wiped that drive, it’s still your id and you can get it back up.

Here’s the post

Replacing the primary drive?

If you want to upgrade your main drive, you should back up the node ID first. But it’s not necessary as Weynand pointed out. But in general, it will keep your older, lower TFT price entry. Check the FAQ we cover this. Q: Where is the 3node ID stored? A: For the current Zero-OS version, the node ID is stored in the first SSD you install on your 3node. If you change or erase this disk, this disk will lose its current 3node ID. Q: Is there a way to backup my node ID in order to restore a 3nod…

[LongStreetTech]

Guide? What is that? Never heard of such a thing. Probably made up by the same people that claim instructions come with products. Appreciate the link. I think it was a striped raid array….I don’t remember. But I’ll read your much appreciated link when I get time to make changes that involve storage on another one. And glad to know the CPU’s can be swapped on that system whenever I get time without worry. Have a good one!

[scott]

The spec is 95% uptime required for DIY farmers, 97% for certified, and now 99.8% for gold certified. That equates to 36, 21.6, and 1.44 hours of allowed downtime per month.

Of course, the situation could still arise and yes the current economically rational thing to do is keep the machine offline until the next month. The biggest reason not to do so is that this hurts your farm’s overall reliability metrics and someone might thus choose to deploy their workloads to a different farm. This is one reason it’s important that we eventually implement some incentive for farmers when their nodes are utilized, along with tools for deployers to view metrics beyond a node’s current uninterrupted uptime.

[Mik]

@scott “along with tools for deployers to view metrics beyond a node’s current uninterrupted uptime.”

excellent point!

What data would be good to have instead of only having the current “uninterrupted uptime”?

Here are some ideas.

• Overall total uptime would be amazing to have.
  • Say during a complete year or more, what is your uptime, not just in terms of by a month by month basis.
• how many times per a given period do you do maintenance, updates
• how long the maintenance windows are?
• % of total uptime without considering the maintenance window, or data showing downtime for maintenance vs downtime for external, non-maintnance issues (connectivity, power outage, etc.

A long maintenance window that happens only once a year can be more interesting than dozens of small maintenance window per month, also different users could prefer one or the other.

Also, downtime caused by optimally scheduled and previously announced maintenance window is a “better” downtime than a downtime during the peak period of the day, caused by, e.g. power outages.

[ParkerS]

I would love a way to report a scheduled downtime, for instance. For example I will be taking my entire farm offline as soon as the parts are here (probably Monday) because I’m putting in new l3 switches, Ned ssds and 400gb of ram.

With my nodes being gateways that downtime could theoretically cause a service interruption or degradation, and honestly I’d just much rather warn people then just pull plugs and haul ass to get back up

[scott]

Nice thoughts on the metrics, @Mik. For a farm, this could be look at average number of outages per node and average duration of outage, over a given time period.

We’ve had some discussion before around how farmers can signal to their deployers about planned maintenance. Maybe this could be built into the system at some point. For now farmers can run websites, email lists, Telegram announcement channels, etc, to serve this purpose. Maybe that’s the best long term solution anyway.

[FLnelson]

There’s a lot of way to calculate reliability for a farm, but I personally don’t think being more complicated than a 3 month rolling average is needed. Had a month where you had to take it offline a lot to configure? Ok, well you’re penalized for not being a network engineer and being slow on the setup, but your bad reliability score will wear off soon.

Overall time since joining the network will also be vital. Older nodes are more trustworthy than new ones.