Skip to content

pool: improve pool.Connect logic #372

Closed
@oleg-jukovec

Description

@oleg-jukovec

The current pool.Connect logic:

  1. Pool will use only endpoints available at the call moment.
  2. If there are no available endpoints then a pool will not be created.

We need to rework this two moments:

  1. Some endpoints may gone forever from a client point of view. We could avoid it and try to reconnect to unavailable endpoints in a background (as it already done if an endpoint became available).
  2. We could create/use (with an error) an empty pool since we have Add and Remove methods. So from the API point of view it would be nice to have an ability to create a pool object and add instances later.

Metadata

Metadata

Assignees

Labels

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions