Auto retry potential DB transient failures with tenacity

[viniarck]

As highlighted on pymongo docs, it's responsibility of the application to handle certain exceptions like AutoReconnect when the driver can't auto retry. So, this issue is to add support for retry capabilities with tenacity for certain writes and reads to make the integration production-grade. If you have suggestions to consider other retry library let me know, tenacity is great, I've used it in production before.

What's being proposed here also goes hand in hand with the dead letter dict structure, mapped on this issue, if all retries fail it'll land there.

Side note: There's also use case for tenacity on other NApps as mapped here, I think it makes sense to ship it on kytos core when we get the chance in the next release.