stacklok / toolhive / 20818260476

50%

push

github

Add retry logic for health check failures (#3222)

Fixes #3221

Previously, remote MCP servers would shut down on the first health
check failure, requiring manual restart. This change adds:

- 3 retry attempts before marking a server unhealthy
- 5 second delay between retries
- Logging for retry attempts and recovery

This prevents transient network issues (like VPN/firewall idle
connection resets) from immediately killing the proxy.

Signed-off-by: Frédéric LE FEURMOU <flfeurmou@indeed.com>

2 of 31 new or added lines in 1 file covered. (6.45%)

28912 of 57809 relevant lines covered (50.01%)

58.14 hits per line