Upgrade UnboundID LDAP SDK to 7.0.3

[ankit--sethi]

The current version is 4 years old; newer versions have substantive improvements to connection pooling and health check behavior.

[elasticsearchmachine]

Pinging @elastic/es-security (Team:Security)

[elasticsearchmachine]

Hi @ankit--sethi, I've created a changelog YAML for you.

[elasticsearchmachine]

Hi @ankit--sethi, I've updated the changelog YAML for you.

[piotrsulkowski-elastic]

LGTM!

Mind that normally I wouldn't approve anything yet, as I'm pretty new, but this change does seem fairly uncontroversial and safe.

I'm actually surprised that it fixes the mentioned issue, but if it does, then great! I imagine that in the most common scenario of single connection being broken, the new strategy to wait until a new connection is created (instead of trying another one from the pool) is slower, however the benefits seem to outweigh that problem. It seems that the main benefit is being able to detect broken servers promptly and return a good connection, as opposed to circulating through different servers, some of them being shut down, erroring each time until they timeout on some arbitrary ttl. From the release notes, it sounds to me more like an accidental improvement, but as you mentioned earlier, it would be difficult to fix it on our side if the library didn't fix it.

[elasticsearchmachine]

Hi @ankit--sethi, I've updated the changelog YAML for you.

[slobodanadamovic]

LGTM 👍

[elasticsearchmachine]

💚 Backport successful

Status	Branch	Result
✅	9.0
✅	9.2
✅	8.19
✅	9.1