To continuously monitor JVS liveness, we need a basic prober. The basic logic is:
Give a JVS API endpoint
Request a token
Validate the token
Write a success metric data point
Repeat after a while
We then set up alerting based on the success metric.
We already have jvsctl that's capable of requesting/validating token. We can build an image on top of it.
We can use log-based metrics.
We can use cloud scheduler for the repeat.
To continuously monitor JVS liveness, we need a basic prober. The basic logic is:
We then set up alerting based on the success metric.
We already have
jvsctl
that's capable of requesting/validating token. We can build an image on top of it. We can use log-based metrics. We can use cloud scheduler for the repeat.