feat: add initial healthcheck endpoint (ampd)

[eloylp]

Description

At Eiger we are working on the Axelar Solana implementation. During this work, we realised ampd doesn't have a defined way to check the liveness of the process from an external perspective. This is specially useful when deploying workloads in i.e Kubernetes, as the cluster can query the process in order to check if its alive or not.

This is a PR proposal for adding an HTTP /status endpoint to the ampd process, that can be used by external systems to determine the liveness of the process.

Closes #12

Todos

• Unit tests
• Manual tests See this comment .
• Documentation
• Connect epics/issues

Steps to Test

Manual

1. Start the ampd server.

2.  $ curl localhost:3000/status

Unit tests

$ cd ampd && cargo test

Expected Behaviour

Once ampd starts its activity, an HTTP request to the /status endpoint should return a 200 OK response in json format:

{
 ok: true
}

Other Notes

This is just a dummy HTTP health check endpoint that only provides a probe of liveness of the process. It doesn't query any subsystem in order to check their internal health. In a future, such option could be revisited.

[eloylp]

Normally i do not like to use new() for doing things like binding sockets. Initially i consider this approach doable taking into account how the rest of the codebase interacts with this server.

[eloylp]

Manual testing

In order to provide certain confidence the healthcheck endpoint is not disturbing other components, the ampd daemon was tested by:

1. Bringing up locally the tofnd (pre-requisite). We used the docker-compose way sharing the port with the host.

2. Configure the ~/.ampd/config.toml file with the following content:

   tm_jsonrpc = "https://axelar-testnet-rpc.qubelabs.io:443"
   tm_grpc = "http://axelar-testnet-grpc.qubelabs.io:9090"
   event_buffer_capacity = 1000
   health_check_bind_addr = "0.0.0.0:3000"
   
   [service_registry]
   cosmwasm_contract="axelar1hrv2zf8xsfsey2umewng68c8pwsyktq4sm8u6k5sch4gqkn0qxrsg28xnx"
   
   [broadcast]
   batch_gas_limit="10000000"
   broadcast_interval="1s"
   chain_id="devnet-amplifier"
   gas_adjustment="2"
   gas_price="0.00005uamplifier"
   queue_cap="1000"
   tx_fetch_interval="600ms"
   tx_fetch_max_retries="10"
   
   [tofnd_config]
   key_uid="axelar"
   party_uid="ampd"
   url="http://127.0.0.1:50051"
   
   [[handlers]]
   type = 'MultisigSigner'
   cosmwasm_contract = 'axelar1lz2hevr93dwa3l86n7hnyy36age39jsk0e8nsj804xr3n8hh958q79m2sq'

3. Execute cd ampd && cargo run :

   ❯ cargo run
       Finished dev [unoptimized + debuginfo] target(s) in 1.19s
        Running `/home/eloylp/projects/eiger/axelar-amplifier/target/debug/ampd`
   2024-04-10T13:15:35.305195Z  INFO ampd: found config file /home/eloylp/.ampd/config.toml
   2024-04-10T13:15:35.308031Z  INFO ampd: starting daemon args=Args { config: ["~/.ampd/config.toml", "config.toml"], state: "~/.ampd/state.json", output: Output::Text, cmd: () }
   2024-04-10T13:15:35.308152Z  INFO ampd::state: loading state from disk
   2024-04-10T13:15:37.850389Z  INFO ampd::health_check: Starting health check server at: 0.0.0.0:3000
   2024-04-10T13:15:40.417094Z  INFO ampd::state: state updated handler="multisig-signer" height=12887365

   We can see the output Starting health check server at: 0.0.0.0:3000 , which announces the healthcheck endpoint should be reachable at this point.

4. Reach the health endpoint:

   ❯ curl localhost:3000/status
   {"ok":true}

[eloylp]

The work of this PR is being upstreamed at axelarnetwork#344 . We are waiting for feedback there. Closing this one.