[gql] Increase EPOCH_DURATION_MS to reduce tests interacting with reconfiguration

[wlmyng]

Description

When we submit a transaction to the test cluster while the nodes are reconfiguring for the new epoch, they will likely time out, causing several tests to frequently flake out.

Test plan

Flakey tests pass without multiple retries

---

Release notes

Check each box that your changes affect. If none of the boxes relate to your changes, release notes aren't required.

For each box you select, include information after the relevant heading that describes the impact of your changes that a user might notice and any actions they must take to implement updates.

• Protocol:
• Nodes (Validators and Full nodes):
• Indexer:
• JSON-RPC:
• GraphQL:
• CLI:
• Rust SDK:
• REST API:

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
sui-docs	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Dec 23, 2024 11:48pm

3 Skipped Deployments

Name	Status	Preview	Comments	Updated (UTC)
multisig-toolkit	⬜️ Ignored (Inspect)	Visit Preview		Dec 23, 2024 11:48pm
sui-kiosk	⬜️ Ignored (Inspect)	Visit Preview		Dec 23, 2024 11:48pm
sui-typescript-docs	⬜️ Ignored (Inspect)	Visit Preview		Dec 23, 2024 11:48pm

[stefan-mysten]

Thanks @wlmyng. No expert here, but I was wondering if it wouldn't make sense to have a higher epoch duration to around 60s, instead of requiring to force_reconfiguration.

Also, there might have been a reason as to why the duration was set to 10s. Do you happen to know why?

[wlmyng]

Thanks @wlmyng. No expert here, but I was wondering if it wouldn't make sense to have a higher epoch duration to around 60s, instead of requiring to force_reconfiguration.

Also, there might have been a reason as to why the duration was set to 10s. Do you happen to know why?

I've actually bumped the epoch duration to 300s, but otherwise some indefinitely long amount of time would work

I think the original idea was that we'd keep EPOCH_DURATION_MS short for testing. But what ends up happening is that transactions end up waiting for reconfiguration, which may even happen back to back, causing transactions to time out. What we should do instead is basically set epoch duration to some indefinitely long amount, and users of TestCluster should instead progress by trigger_configuration.

Note that this only affects the TestCluster instantiation in sui-graphql-rpc, and doesn't change the defaults from the test cluster crate. In a follow-up, I plan to expose a way to adjust this per instantiation in the graphql crate.

And I think it was set to 10s basically since its conception. Think the last relevant change was from 14 months ago, from me (ignoring Brandon's pg refactor)