Change altruistic service to use nodes directly instead of relaying

[nymd]

Currently the altruism service works by relaying the fallback calls to a special set of Pocket nodes. These nodes, run by PNI, are out of date and only used for altruism. The feature they use to do this is not supported in future versions and it is an inefficient relay system by all measures.

The altruistic node service should be replaced with service that directly calls the PNF hosted blockchain nodes in US-East-2.

There is an internal ALB that is used for node communication located in US-East-2.

All VPCs where the Portal API exists should be VPC peered with the node-infra-manual VPC in US-East-2: https://aws.amazon.com/premiumsupport/knowledge-center/elb-access-load-balancer-vpc-peering/

The portal API should keep a stored list of the chain's full URL for use in the backup service.

That stored list could be a GitHub secret that is loaded into an env variable and parsed or the backup location could be stored in the mongo DB as part of the blockchain record.

[nymd]

As raised by BenVan here: https://forum.pokt.network/t/pushing-altruistic-servers-to-the-edges/1053 the altruistic service could be distributed between PNF's nodes and other nodes run by our bootstrapping partners.

An array of altruists for each chain could be kept and the altruist randomly chosen from the available.

This might be better done by deploying nginx or using an ALB as the altruistic front so that uptime can be automatically monitored.

[nymd]

Closed with #100