This is open-orchestrator.eth - we’ve just launched our official campaign! Excited to officially join the community.
We run a globally distributed orchestrator which is completely open source - we built our stack on Kubernetes (a popular solution for managing complex software architectures) and open sourced the entire codebase for the community.
Our orchestrator will be managed using “OIPs” - whenever a protocol-level change is being considered (e.g. reward/fee cuts), or significant change to the off-chain orchestrator parameters (e.g. price per pixel, max face value for tickets, etc), an Orchestrator Improvement Proposal (OIP) must be submitted by the maintainer(s) or by a delegator. This will consist of a PR similar to OIP-1 which outlines the proposed changes and the reasoning for the change. The community will be given seven (7) days to leave comments on the PR before it is merged and executed. If you are a delegator, please turn notifications on for PRs, so you can participate in these discussions. We want complete transparency with our delegators to make sure we maintain the ideals of this great ecosystem.
I also work at Livepeer on the protocol design and web3 initiatives in the company. This is a personal project (not sponsored by Livepeer) built so that it can give back to the community by giving an open source example of one way a globally performant orchestrator can be set up. You’ll see me in the Discord channels as 0xcadams.
Feel free to reach out to me on Discord, I’m always available to help set up new orchestrators, especially using Helm/Kubernetes!
The open orchestrator experienced a temporary outage from Nov 16th at 21:50:30 UTC to the following morning at 9:15:00 UTC. During this time, the Arbitrum nginx proxy got into an unrecoverable state on all clusters, despite the upstream Arbitrum failover nginx continuing to respond appropriately. This was likely due to a temporary outage in the community Arbitrum node (as has been seen in the past), which triggers the nginx pod to cache/continue to return a bad response code.
There has been a bug filed in the repo and it will be fixed ASAP. Please subscribe to the Github issue for updates.
We will be winding down our Frankfurt region for the foreseeable future to cut down on costs. The Virginia region consistently receives more traffic and has good enough networking to provide a comparable service without the expense of the Frankfurt region (near $400-500/mo currently). Once it becomes more economical to operate again, we will spin the Frankfurt region back up.
We regret to inform the community that Open Orchestrator will be shutting down due to financial conditions and unsustainable infrastructure costs. This was a difficult decision to make and we are deeply sorry for the inconvenience it may cause.
We have been a part of the Orchestrator community for nearly eight months and we have enjoyed every moment. We understand that this news may come as a surprise to our delegators, and we want to thank you for your loyalty and support.
Since our service will no longer be available, we recommend that you find an alternative orchestrator(s) to meet your needs. We apologize for any disruption this announcement may cause and we wish you the best of luck in finding a suitable replacement.
The source code will remain available indefinitely for anyone looking to spin up a cluster with Kubernetes. We will be spinning down operations on Feb 13, 2023.
Thanks for sharing your Kubernetes setup.