Passively Test F3 on Mainnet

[masih]

The following task list outlines the milestones required to complete the passive F3 testing on mainnet.

See F3 Mainnet Passive Test Plan for more details.

• Testing utilities for passive testing in network deployment #229
• Comprehensive Simulation Testing #226
• Monitoring Infrastructure Setup #227
• Local Testnet Deployment #228
• Calibnet Deployment: Deploy the F3 protocol on Calibnet to mimic the conditions and processes of a mainnet rollout.
• Post-Calibnet Deployment Metrics Analysis: Analyze the collected metrics for performance and correctness indicators. Confirm that all systems function as expected based on the metrics.
• Potential Pre-mainnet F3 Implementation Adjustments: Based on insights from the Calibnet deployment, make any necessary adjustments. Freeze the F3 Alpha implementation in preparation for mainnet deployment.
• Mainnet Deployment: Roll out the optimized F3 implementation across the mainnet. Follow a phased approach to monitor performance and impact during the rollout.
• Post-Mainnet Deployment Metrics Analysis: Continuously monitor the F3 metrics post-mainnet deployment. Assess performance, correctness, and uptake to determine if further adjustments are needed.

[masih]

This was intented to be the top level umbrella issue. But now it's in Milestone 1.

If that's where we want it to be then we should remove milestone 0 tasks from it.
CC @Kubuxu @jennijuju

[rjan90]

Goal is to get up to 1000, but unclear if it is possible to get that number with the rollout of v0.2.0

[Stebalien]

We're going to punt the first round of testing till tomorrow to give the pubsub mesh some time to form (needed to reboot the bootstrappers to join the new network).

[rjan90]

@masih Can we close this ticket, and instead open smaller testing tickets for individual passive testing efforts?

[masih]

Can we close this ticket, and instead open smaller testing tickets for individual passive testing efforts?

My recommendation is to indeed open smaller tickets for explicit test efforts, and reference them under this one. Because: 1) this is the ticket that has been referenced in previous community posts. 2) it gives the community a single point of reference into what actually gets tested in passive testing along with the ability to drill down further should they want to learn.

[rjan90]

• Phased approach - run it in a few steps
  • Thursday/Friday (2024-11-21 & 22) get back into tooling, deploy with a few percentage of the network.
  • Monday (2024-11-25) start with 10 and then ramp up to 30, 50 and 100 percent.

[rjan90]

This is what happened last week

• Last Friday we got back into tooling, and deployed passive testing with around 5 MinerIDs participating in the passive testing. It ran successfully over the weekend without any hiccups.

This is the plan for this week

• Today (2024-11-25) we are planning to bump the participation to 100 MinerIDs, then we will continue to bump it towards 200 MinerIDs joining passive testing, 600, 1000 and everyone (approximately 2000 MinerIDs)

[BigLep]

I'm posting a few things here as I believe f3 team members have dropped off for the day.

1. For 2024-11-26, status/progress updates will be posted in F3 passive testing lotus#12287 since that has a longer history.
2. I believe that passive testing got up to 200 minerIds today (2024-11-25) but was paused around 20UTC. More passive testing will happe 2024-11-26.
3. These are the relevant dashboards:
   Calibration: https://grafana.f3.eng.filoz.org/public/calibnet
   Mainnet: https://grafana.f3.eng.filoz.org/public/mainnet
4. A relevant blogpost for this passive testing work was published here: https://medium.com/@filoz/finality-unveiled-passive-testing-to-mainnet-launch-of-f3-fast-finality-03e09bc68de5