v0.4.0

This release introduce a new concept regarding capacity reservation, the capacity pool. This goal of this new feature is to decouple de reservation of capacity from the workload provisioning.

With the capacity pool, you can now provision and cancel workload as much as you without payment required. Instead, the workloads are linked to a capacity pool from which they consume cloud units.

It also allow to extend the lifetime to existing workload by just topping up the linked capacity pool.

You can find more detail information in the forum announce post and the complete specification and rational behind this at https://github.com/threefoldtech/tfexplorer/blob/master/specs/modify_IT_contract_over_time.md#extension-of-the-capacity-pool

Changelog

• #58 Add node and farm iterator in client
• Show unavailable capacity as grey in frontend
• #95 tfuser: fix network add-node not updating schema with network resources
• #74 Don't return signature field in user listing
• Add yggdrasill range in the list of private network to avoid the network generation to use yggdrasil IP address as public endpoint
• #97 tfuser always create a bcdb client
• update docs and extend makefile to generate docs
• #116 Validate zdb, container and volume disk type
• Change the way the workloads are sign. Drop the idea of having the full reservation json encoded as signature challenge. Instead each workloads needs to implement a function that return a hash this hash is then used to create the signature
• Remove the concept of provision expiration for workloads. The user is now fully in control of when a workload is deleted even in the case of a node not provisioning it. Like this the user can send some provision request and they will eventually be fulfilled by the node
• create a new versionned path for the API. the API is now exposed at /api/v1 instead of /explorer
• only use online nodes to compute amount of resource units show on home page
• #141 change how delete and provision signature are accepted and verified. After this commit, user that are requested in the delete or provision consensus are only able to push a signature once. If they already have a signature registered and try to put a new one, an 400 bad request error will be returned by the API
• #144 Mark workload that are not allow to be deployed as invalid to avoid them to be picked up by provisiond on next restart
• Remove panel for total number of nodes and gateways in frontend
• #148 also take in account the gateways when checking for free to use node during pool creation
• 146 Buffer paid capacity channel
• add field to container workload to allocate yggdrasil IP