Pool issue, can't get healthy conn

[RyouZhang]

When upgrade to v9.16.0, we get many error log like this

redis: 2025/10/28 03:26:44 pool.go:449: redis: connection pool: was not able to get a healthy connection after 3 attempts

Then we rollback to v9.14.1, no any error log.

[ndyakov]

Hello @RyouZhang , we added some additional logging to help us improve the pool in the near future, this would be expected if you have set max lifetime of the connections. Can you share your configuration and setup?

[RyouZhang]

Config like this:

PoolSize: 32,
MinIdleConns: 4,
ConnMaxLifetime: 300s,
MaxRetries: -1,
ConnMaxIdleTime: 30s,
PoolFIFO: true,
Protocol: 2,

[ndyakov]

@RyouZhang this makes sense, My assumption would be that you are getting those messages when the connections reach their maxlifetime and the pool considers them "unhealthy" so they can be changes. It should be safe to use 9.16, feel free to observe and we are planning to address the logging (have different levels, so you can silence those) in the near future. For now I will take a note and silence this log in the future release.

[RyouZhang]

ok, I know, maybe need to change log message and use warn level

[ndyakov]

@RyouZhang sadly, the library doesn't support different log levels at the moment, we are planning this for v10.

[RyouZhang]

pls change content like
conn be removed, because trigger maxlifetime; now we can't detect the real unhealth

[yuwenlong]

We also encountered this problem.