Update version of prometheus postgres exporter in neon compute image to avoid conversion error for replication_slot

[Bodobolero]

Steps to reproduce

https://neonprod.grafana.net/d/000000039/neon-compute-metrics-by-endpoint-id?var-endpoi[…]qm&var-env=dev&orgId=1&from=1709527200000&to=1709527800000
](https://neonprod.grafana.net/d/000000039/neon-compute-metrics-by-endpoint-id?var-endpoint_id=ep-sparkling-glitter-w16jf9qm&var-env=dev&orgId=1&from=1709527200000&to=1709527800000)

https://neondb.slack.com/archives/C04DGM6SMTM/p1709551068646819?thread_ts=1708462863.991469&cid=C04DGM6SMTM

Expected result

No error during metrics scraping on Neon compute nodes

Actual result

The following error in compute log

ts=2024-03-04T04:49:59.694Z caller=collector.go:194 level=error msg="collector failed" name=replication_slot duration_seconds=0.001089597 err="sql: Scan error on column index 3, name \"active\": converting driver.Value type bool (\"true\") to a int64: invalid syntax"

prometheus-community/postgres_exporter#769

I suggest to upgrade postgres_exporter to then current version or a version later than the one fixing the bug.

It seems we currently use

root@neonvm:/# postgres_exporter --version
postgres_exporter, version 0.12.0 (branch: HEAD, revision: f9a1edbbc6d629663aa658378e1d989293e6e488)
  build user:       root@463311dfdfb6
  build date:       20230321-23:52:03
  go version:       go1.20.2
  platform:         linux/amd64
  tags:             netgo static_build

Environment

Logs, links

neon/vm-image-spec.yaml

Line 179 in 3114be0

FROM quay.io/prometheuscommunity/postgres-exporter:v0.12.0 AS postgres-exporter