RFC: Adjustment of Altair collator rewards vol. 2

Hello Altair community!

I am the original author of the proposal that helped Altair collators to achieve a sustainable level of rewards that now helps them to maintain a high-quality node operator service. A part of this proposal was a 3-month periodic evaluation of the AIR reward so it matches the desired value of $125 that was agreed on in the proposal mentioned above.

Unfortunately, since then we have seen a significant downturn in markets which led collator rewards to deteriorate and again begin unsustainable. So, in the spirit of the original proposal, I am suggesting a monthly reward to be 15,122 AIR.

This is based on the following calculation

  1. 3-month average being $0.008266 (source - TradingView, AIR/USDT pair on gate.io which is the most liquid one)
  2. Desired reward is $125

Therefore ~15,122 AIR should be the desired reward now. I will appreciate any thoughts or comments on this proposal, thank you :slight_smile:


I made mention of this a few weeks ago (RFC: Adjustment of Altair collator rewards - #19 by TheMarcus) so I’m glad we’re revisiting after 3 months from the last adjustement.

I concur with @pmensik that an adjustment is needed seeing as the price of AIR has fallen lower since our first adjustment. Recommend on-chain proposal.

1 Like

@Rhano Can we get an update from the implementation of Block rewards for Altair before this discussion on the adjustment for the collator rewards keeps going on?

1 Like

Hi @Tjure07

The Block Rewards and Collator Cycle have been successfully implemented on Altair:

  1. Council Motion 102: Runtime Upgrade 1027
  2. You can find forum posts about upcoming upgrades here:
    Upcoming Runtime Upgrades on Centrifuge 1019 & Altair 1027

Councillors are no anymore involved in Collators payments because the funds came directly from Treasury.
In case Collators would like to increase Collators rewards payments they should initiate a separate Runtime Upgrade in order to update the value on-chain. That will require SuperMajorityApproval and all Altair Community should vote on this proposal.


Good day @pmensik

Thank you for your proposal.
Can you please describe the actual cost for the collator for running the Altair node?

Hey @ImdioR,

we already discussed that in an original proposal and $125 was an average we came to through the poll that was ran among the collators. As far as Polkadotters costs go, here is the summary

Collator costs

  • bare metal on (currently on Ikoula , Agile XL Pro v2) - $80
  • 1 hour - maintaining the collator (usually $50 in crypto for sysadmin salary)
  • 25$ - flat fee for stuff like running Grafana, Prometheus, alerts and other services related to the safe operation of a collator

Not currently running a backup since there’s not really any activity on the chain but I plan to do so in the future - so that would be another $50.

So, without a backup, that accounts for $155/month so let’s round it to $200 to count in some actual profit for running the service.

1 Like

Good day @pmensik

Agile XL Pro v2
Intel® Core i9 9900K @3.6 GHz (8C/16T) 64 GB DDR4 1 TB SSD NVMe Intel Graphics UHD 630

How many nodes are you actually running on one dedicated server?

At the moment it’s two due to most of the chains not being profitable now and there’s not much happening on Altair - however, our standard operation is 1 collator per 1 dedicated machine as it’s in the documentation. Also, most of the parachain dev teams are recommend running nodes at least matching the spec so that’s why we follow it.

So the cost should be divided? In this case why we are calculating the price of a dedicated server for 1 chain if more nodes are runned on the same server?

Altair Collator Requirements

Instance Resources

  • At least 4vCPU
  • At least 16GB RAM
  • At least 500GB SSD
**Agile XL Pro v2**
Intel® Core i9 9900K **@3.6 GHz** (8C/16T) **64 GB** DDR4 **1 TB** SSD NVMe Intel Graphics UHD 630

Running this dedicated server with such high performance not sure how this is useful for the Altair chain and what is the benefit for all of Altair Community to finance this.
Also please note that another Altair Collator already pointed this out in prev topic here and here

We are operating on AMD Ryzen 5 3600, It uses avg. 6% of cpu, less than 1% of I/O nvme and 5% of the memory. It’s a good machine for parachains where also little performance is required. Also in the future we will use relay-chain from rpc where the parachain can even have more capacity to run.

I will explain my point of view. Before the Altair Councillor decided to approve or disapprove the Treasury proposal, but now this is regarding all Altair Community, of which I’m a token holder.
In my personal opinion running i9-9900K with 64 GB is not necessary for Altair node (check the specs suggested here ) + Collators hardly run only 1 node on a dedicated server like this.
So invoicing the FULL cost of a dedicated server only on Altair Community is not fair in my opinion.

First of all, I would address the vCPU and its pricing. I think it’s quite obvious that there has to be a trade-off for such a low price, actually, there are two.

Reliability - as professional node operators, we have to maintain a high reputation and trust. The only way to achieve it is to either run your own hardware (we are doing it too) or rent dedicated servers which you have full control of. In case you are running VPS, you are exposing yourself to the risk of platform and our experience so far is that popular providers such as Contabo (which is the cheapest one) has an outage at least once a month. Plus there are various benchmarks showing that the performance of their SSD drives is not on par to what they are claiming. So either we provide a quality service or we chase the lowest price, our choice is the former.
Decentralization - the very reason why we are allowing collators to be in the active set is decentralization, more liveness and censorship protection. If you start to encourage collators to run on a cheap VPS, they can very well end up in the same data center - and in case of an outage, the whole network can go down or at least experience slower block production. Regarding decentralization, we should rather encourage a multitude of geo and data center locations to ensure the liveness of the network.

As far as the costs go - to put it simply, we pay $80 for the server, $25 is our standard fee for managing all the monitoring and alerts, $25-50 is a backup (which we don’t have now as stated above) and say 1-hour in a month for maintenance which is $50. So even if you divide the server cost by two, it gives us 40 + 25 + 50 which is $115. We are asking for $125 monthly which gives us a mere profit of $10 every month - and I bet I spend more than 1-hour taking care of all of the collator tasks and activities around Altair and Centrifuge.

As you might noticed, we are usually the first ones to respond, to check whether upgrades are correct and afaik we haven’t had any downtimes so far - so in this context I find the reward more than reasonable. But I definitely wonder what other collators has to say about this as well :slight_smile:

I doubt that @Ilhan nodes went down or were kicked off from the collator list due to downtimes.


:raised_hands:Thank you for your answer.

I think @pmensik states it very well in his last reply. You either choose high quality or low cost.

We have opted for low cost, but at the risks pmensik mentioned. We certainly don’t want to be a weak link in the proverbial chain, but for now it works for us and seems to work alright for the network. However, should the Altair network ever gain more activity, I’m certain our provisions would no longer be adequate.

I would be curious to hear what the Centrifuge/Altair senior team members think about this topic. What do they prefer / expect from those helping to operate / secure the networks which represent the projects they are building?

So, without a backup, that accounts for $155/month so let’s round it to $200 to count in some actual profit for running the service.

I would like to address the last bit - yes, it was $155 because the Altair collator was running on a single server and we didn’t know at the time that the price will drop and it would force us to run multiple collators on one machine. Also, I round it to include a small profit for the service itself - but I actually proposed $125 to be paid every month so I even discounted it when making the proposal :slight_smile:

Although I am likely not technically competent enough to put forth specific assertions about what we should and should not be paid, I have read through the comments and understand both sides of the argument.

Here at Lucky Friday, we run our own bare metal servers in SOC2 data compliant centers. We use newer NVMe machines (although our Altair may be on an older SSD machine; I’d have to check with Will/Paradox) and have also made a commitment to carbon neutrality, all of which drives up the costs. While I cannot claim to know how much those are exactly, I believe the current rewards payment from the agreed proposal would cover said costs. Moreover, as noted by others, there’s not a ton of activity on the chain and therefore maintenance on this node is not high.

Ultimately it will be the Altair community who will pay for this because rewards will be paid from treasury funds, but we have yet to hear of anyone (beyond collator runners and team members) chime in on this. I will be sure to ask the rest of the LF team for their thoughts and return back with additional commentary. Perhaps we could also discuss this on Wednesday during the upcoming governance meeting.

@pmensik and other collators on Altair and Centrifuge; please check this post where Lucas has started a discussion around criteria for Centrifuge and Altair collators - this could also be useful to have in mind in relation to the financial aspect of this proposal.

cc. @The_Phunky_One_Lucky @TheMarcus @Ilhan @n1trog3n @PathrockNetwork @Ryabina @y3v63n @Su-Staker.Space @shez @Validatrium @Denys_Validatrium @Igor_Validatrium @cardinate


@Rhano thanks for the mention. I do think focusing on a combined collator program makes sense for the reasons stated here!

Just saw Lucas’ post in the Discord server and left some comments!

1 Like