Skip to main content

Collator Requirements

How to become a collator

Semi-Permissionless collator

To become a collator on our networks, you need to be approved by governance (2/3 MainCouncil or referendum).

  1. Apply for candidacy by reserving the bond (Extrinsics - CollatorSelection - Apply for candidacy)
  2. Request for approval from governance.
  3. Once governance proposal is executed (Extrinsics - CollatorSelection - Approve application). The onboading takes n+1 session.

Collator staking requirements

  • Bond: 3,200,000 ASTR tokens (3.2M $ASTR)
  • Meet hardware requirements

If your node stops producing blocks for 1 session, your node will be kicked out of the active set and 1% of the bonded funds will be slashed. Running a node with low performance can lead to skipping blocks which may result in being kicked out of the active set.

tip

If you want to withdraw the application and unlock the bond Extrinsics - CollatorSelection - Close Application


System requirements

A collator deploys its node on a remote server. You can choose your preferred provider for dedicated servers and operating system. Generally speaking, we recommand you to select a provider/server in your region, this will increase decentralization of the network. You can choose your preferred operating system, though we highly recommend Linux.

note

Under async backing, collators must not only build blocks quickly, but also keep up with relay-chain updates, branch churn, candidate generation, and candidate distribution. In practice, this means collator performance depends on CPU quality, dedicated core count, RAM, storage latency, and network stability, not only single-core frequency.

Hardware requirements

The configurations below should be treated as the minimum production baseline for collators. Machines below these specifications may fall behind during syncing, reorg-heavy periods, or candidate propagation, which can reduce canonical authorship competitiveness and increase the risk of being removed from the active set.

caution

Make sure your server is a bare metal only dedicated to the collator node, any unnecessary other process running on it will significantly decrease the collator performance. We strongly discourage using a VPS to run a collator because of their low performances.

Collators are the nodes which require the most powerful and fast machine, because they only have a very short time frame to assemble a block and collate it to the relay chain. To run a collator, it is absolutely necessary to use a CPU of minimum 4 Ghz per core (strong single-core performance strongly recommended) and a NVMe SSD disk (SATA SSD are not suitable for collators because they are too slow).

ComponentRequirement
SystemUbuntu 22.04 or 24.04 LTS
CPU12 dedicated modern x86-64 cores - minimum 4 Ghz per core
Memory32 GB
Hard Disk1 TB SSD NVMe, low-latency SSD
Network500 Mbit/s symmetric, Stable broadband with good latency
tip

Shibuya is the perfect network to test out your knowledge about running nodes in the Astar ecosystem. To join the collator set on Shibuya you need to apply for a 32k SBY fund. If you never operated a collator node, we strongly encourage you to spin up a Shibuya collator node to start before thinking about mainnet. A perfect start is our secure setup guide.