The case of the disappearing GCP metadata-server

I was browsing the logs as the service starts, as I want to ensure it uses GCP for backups, and I saw:

app-analytics-db-0 postgres 2020-12-20 14:16:08,259 - bootstrapping - INFO - Figuring out my environment (Google? AWS? Openstack? Local?)
app-analytics-db-0 postgres 2020-12-20 14:16:08,264 - bootstrapping - INFO - Could not connect to 169.254.169.254, assuming local Docker setup
app-analytics-db-0 postgres 2020-12-20 14:16:08,265 - bootstrapping - INFO - No meta-data available for this provider
app-analytics-db-0 postgres 2020-12-20 14:16:08,265 - bootstrapping - INFO - Looks like your running local
app-analytics-db-0 postgres 2020-12-20 14:16:08,299 - bootstrapping - INFO - Configuring bootstrap

However;

$ k exec app-analytics-db-0 -c postgres -- curl -v --silent http://169.254.169.254
* Rebuilt URL to: http://169.254.169.254/
*   Trying 169.254.169.254...
* TCP_NODELAY set
* Connected to 169.254.169.254 (169.254.169.254) port 80 (#0)
> GET / HTTP/1.1
> Host: 169.254.169.254
> User-Agent: curl/7.58.0
> Accept: */*
>
computeMetadata/
< HTTP/1.1 200 OK
< content-type: application/text
< metadata-flavor: Google
< server: envoy
< date: Sun, 20 Dec 2020 14:18:23 GMT
< content-length: 17
< x-envoy-upstream-service-time: 0
<
{ [17 bytes data]
* Connection #0 to host 169.254.169.254 left intact

The server is there and responding. Why would Spilo say otherwise?

[CyberDem0n]

wal_gs_bucket should set WAL_GS_BUCKET and from there Spilo builds the WALE_GS_PREFIX: https://github.com/zalando/spilo/blob/master/postgres-appliance/scripts/configure_spilo.py#L731-L749

In case if WALE_GS_PREFIX happened to be set as environment, Spilo will use it directly

[haf]

I’m closing because there’s no negative effect of the reported issue.