Crashloop while running as sidecar for initial minutes

[sijuu]

I am running cosmos-validator-watcher as sidecar with my cosmos node. The cosmos-validator-watcher container crashloops till the node is synced with following error.

11:00PM WRN node http://localhost:26657/ is catching up at block 2688992
11:00PM WRN connected to http://localhost:26657/ (but node is catching up) chainID=athens_7001-1 height=2688992
11:00PM ERR error="no nodes synced"

This causes panic as the container is crashlooping. When a pod restarts, its understandable that it will take some time for the node to be synced. It would be better if cosmos-validator has a while condition check ( as a flag atleast ) which checks for the node to be available and then start the watcher rather than just crashlooping.

[MattKetmo]

Hello, yes indeed, this watcher will fail at startup if no node is up to date in the 5s after launch.

One question though is why you want to run as sidecar instead of running as separate deployment?

The main goal of the watcher is mainly watching validator status on-chain but using several endpoints so it doesn't rely only on 1 rpc endpoint.