chore: lower time for fuzz/race tests

[erikburt]

Lowers the runtime of CI Core fuzz/race tests during executions triggered by PRs.

Changes

Race Tests:

• Lower timeouts to 15s, and count to 5 (on PRs only, still at 10m/50 on scheduled runs)
• Add build cache

Fuzz Tests:

• Lower fuzz timeout to 4 minutes on PRs, 10m on scheduled runs
• Calculating --seconds parameter passed to python fuzz script, based on FUZZ_TIMEOUT_MINUTES
• Better calculation of time elapsed, and individual fuzzer runtimes in python script

Testing

https://github.com/smartcontractkit/chainlink/actions/runs/11734083428?pr=15141

• Core Integ Tests: 2m25s
• Deployment Tests: 2m32s
• Race Tests: 4m39s
• Core Tests: 5m34s
• Fuzz Tests: 5m55s

In best case scenario, long pole is now "Clean Go Tidy & Generate"

---

RE-3205

[jmank88]

Looks good, but could be a little faster? 🤷

[erikburt]

Looks good, but could be a little faster? 🤷

Ok I've lowered them even further. I was having trouble lowering race tests at all, seems no matter what it sits around 8 minutes runtime. I'm thinking it might be build time that is the majority of the runtime.

[jmank88]

Looks good, but could be a little faster? 🤷

Ok I've lowered them even further. I was having trouble lowering race tests at all, seems no matter what it sits around 8 minutes runtime. I'm thinking it might be build time that is the majority of the runtime.

You could try dropping the -json flag from the race tests. I don't know why that is enabled, but it creates extra noise by enabling -v. Might not matter though 🤷

[erikburt]

You could try dropping the -json flag from the race tests. I don't know why that is enabled, but it creates extra noise by enabling -v. Might not matter though 🤷

That seem to have helped by about 30 seconds. Down to 7.5min execution on the race test.

[erikburt]

Ok the build cache does help - down to 5m10s...

https://github.com/smartcontractkit/chainlink/actions/runs/11732567357?pr=15141