Skip to content

BM-3029: Zeroecco/testnet down#2029

Open
zeroecco wants to merge 9 commits into
mainfrom
zeroecco/testnet_down
Open

BM-3029: Zeroecco/testnet down#2029
zeroecco wants to merge 9 commits into
mainfrom
zeroecco/testnet_down

Conversation

@zeroecco

@zeroecco zeroecco commented Jun 2, 2026

Copy link
Copy Markdown
Contributor

Summary

Tears down the production testnet deployments and moves the bare-metal prover Ansible CI onto Tailscale. As part of the cleanup, host-provisioning Ansible roles that are no longer run from this repo are removed, the prover deployment pipeline is simplified, and CloudWatch alarming is repointed at the remaining prover nodes.

Changes

  1. Prod testnet teardown

  2. Removed Pulumi stack configs for the retired prod testnet (Sepolia 11155111 / Base Sepolia 84532) deployments across distributor, indexer, order-generator, order-stream, and slasher.

  3. Removed the matching prod-testnet broker configs (configs/broker/prod-testnet-nightly/, prod-mainnet-nightly/).
    Dropped the l-indexer pipeline and LaunchDefaultPipeline testnet wiring.
    Prover Ansible CI → Tailscale

  4. l-prover-ansible pipeline now joins the tailnet (userspace tailscaled + SOCKS5 proxy) and reaches provers over MagicDNS instead of public IPs; SSH key, inventory, and Tailscale auth key are pulled from AWS Secrets Manager.
    Removed the dead DeployNightly stage and the references to the deleted monitoring.yml/security.yml playbooks; pipeline is now Source → DeployStaging → DeployProduction.
    inventory.yml is no longer tracked (gitignored; real values live in Secrets Manager).
    Ansible slimming

  5. Removed awscli, docker, nvidia, ufw, and vector roles plus monitoring.yml/security.yml — hosts are now expected to be pre-provisioned (incl. metrics shipping handled outside this repo). prover.yml only runs the prover role.
    Alarming

  6. cw-monitoring staging/prod node lists repointed to the current provers (prover-staging, prover-prod) and the retired nightly prover alarm node removed.
    Broker config

  7. Staging broker tuning: max_collateral 5 → 100, peak_prove_khz 80 → 11000, added Sepolia 11155111 override; release-config example paths updated.

zeroecco and others added 7 commits May 27, 2026 11:04
- l-prover-ansible: drop dead DeployNightly stage (no nightly group after
  testnet teardown), remove deleted monitoring.yml/security.yml playbook
  calls, fix secret-name references in docstring
- INVENTORY.md: correct group -> MagicDNS host mapping (prover-staging,
  prover-prod) to match the real inventory
- cw-monitoring: point staging/prod prover alarms at the new hosts and
  remove the retired nightly prover node

Co-authored-by: Cursor <cursoragent@cursor.com>
@zeroecco
zeroecco marked this pull request as ready for review June 2, 2026 22:31
@github-actions github-actions Bot changed the title Zeroecco/testnet down BM-3028: Zeroecco/testnet down Jun 2, 2026
@linear

linear Bot commented Jun 2, 2026

Copy link
Copy Markdown

BM-3028

BM-3029

@github-actions github-actions Bot changed the title BM-3028: Zeroecco/testnet down BM-3029: Zeroecco/testnet down Jun 2, 2026

@willpote willpote left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good aside from the toml changes which I don't fully understand, lets chat offline about it

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants