Open cjcaj opened 6 years ago
There are container lifecycle hooks that we can use to get the enode and write it to efs. https://kubernetes.io/docs/concepts/containers/container-lifecycle-hooks/
It runs the command when the container is created so we'll need to make sure that swarm is running.
Note: create a service for swarm-0 using statefulset.kubernetes.io/pod-name
in k8s v1.9.
@cjcaj 1.9 alpha is out maybe we can try and use that? Also, is there a way we can start monitoring requests to bzz.betwork?
I'm worried that we'll have to recreate the cluster if the update fails somehow. I'll try to see if we can set up the bootnode another way using stable ids.
Regarding monitoring requests, we have this for hma Is this what you had in mind?
Yea something like that for bzz.network would be great.
Adding links to kops issues concerning upgrade to 1.9 https://github.com/kubernetes/kops/issues/5002 https://github.com/kubernetes/kops/issues/4986
NGINX metrics are up. Here's a link to our wiki on how to access it https://github.com/async-la/bzz.network/wiki/NGINX-VTS-Metrics
These new options on POC3 look interesting
--metrics Enable metrics collection and reporting
--metrics.influxdb.export Enable metrics export/push to an external InfluxDB database
--metrics.influxdb.endpoint value Metrics InfluxDB endpoint (default: "http://127.0.0.1:8086")
--metrics.influxdb.database value Metrics InfluxDB database (default: "metrics")
--metrics.influxdb.username value Metrics InfluxDB username
--metrics.influxdb.password value Metrics InfluxDB password
--metrics.influxdb.host.tag host Metrics InfluxDB host tag attached to all measurements (default: "localhost")
I'm wondering if we can investigate what it'll take to have a script execute on each pod boot that stores the enode of that pod somewhere where we can reference.