Optimize multiple agents per build usage by indy CI on Hyperledger Jenkins

Our pipelines for indy core have several stages each processed on Jenkins agents which are one-time on-demand OpenStack minions on Hyperledger Jenkins.

As long as new agent launching is quite slow operation need to:

• investigate how it actually happens on Hyperledger Jenkins for our case
• optimize

Optimization options:

• optimize pipelines according to Hyperledger Jenkins specific
• ask Hyperledger Jenkins admins to make number of hyp-x agents always available
• ask Hyperledger Jenkins admins to allow reuse of launched agents