Skip to content

[kyoto] tighten supervisor restart threshold to 15#489

Merged
0xsiddharthks merged 2 commits intosiddharth/kyoto-upgradefrom
siddharth/kyoto-supervisor-calibration
Apr 27, 2026
Merged

[kyoto] tighten supervisor restart threshold to 15#489
0xsiddharthks merged 2 commits intosiddharth/kyoto-upgradefrom
siddharth/kyoto-supervisor-calibration

Conversation

@0xsiddharthks
Copy link
Copy Markdown
Contributor

@0xsiddharthks 0xsiddharthks commented Apr 26, 2026

Lower KYOTO_MAX_CONSECUTIVE_FAILURES from 30 to 15 so a Kubernetes pod rotation that swaps every whitelisted-peer IP recovers in tens of seconds rather than minutes.

Lower KYOTO_MAX_CONSECUTIVE_FAILURES from 30 to 15. In bip157 >= 0.5.0,
TrustedPeer::from_hostname is consumed on use — popped from the
whitelist when first tried, never reinstated. Once the resolved-IP set
for a hostname is exhausted, only a fresh Builder::build() will run
lookup_host again, and our connectivity supervisor's rebuild path is
that fresh build.

At the previous threshold of 30 consecutive failures, a Kubernetes pod
rotation that swaps the IPs of every whitelisted peer left us
unavailable for several minutes while we burned through the cached
resolution set 3x before rebuilding. 15 halves the worst-case recovery
without churning during a normal single-peer flap. We can re-tune
later from the kyoto_consecutive_failures and kyoto_restarts metrics.
@0xsiddharthks 0xsiddharthks marked this pull request as ready for review April 27, 2026 22:06
@0xsiddharthks 0xsiddharthks merged commit a78332b into siddharth/kyoto-upgrade Apr 27, 2026
4 of 5 checks passed
@0xsiddharthks 0xsiddharthks deleted the siddharth/kyoto-supervisor-calibration branch April 27, 2026 22:07
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.

1 participant