fix(backend): add PORT env to deployments + BEECEPTOR_URL config

16 of 20 rust services were crashing on boot because their main.rs calls
std::env::var('PORT').expect(...). This commit adds the missing PORT env
to each deployment YAML, matching its containerPort (9100-9118).

Also adds BEECEPTOR_URL to the ConfigMap - the payments service requires
it for the mock payment gateway integration.

Adds apps/registry/ with retention script + CronJob (keep last 3 SHA
builds, preserve -latest aliases) to prevent the single-build wipeout
that caused the original registry outage.

AI assistant image also rebuilt: 2876b45 (main branch) - it was on a
ghost SHA tag that was GC'd.
This commit is contained in:
Ashwin Kumar Sivakumar 2026-06-10 23:39:47 +05:30
parent 4034c413c6
commit 827477ac3f
3 changed files with 116 additions and 0 deletions

View file

@ -0,0 +1,6 @@
apiVersion: kustomize.config.k8s.io/v1beta1
kind: Kustomization
resources:
- retention-script.yaml
- retention-cronjob.yaml
namespace: registry

View file

@ -0,0 +1,33 @@
apiVersion: batch/v1
kind: CronJob
metadata:
name: registry-keep-last-3-builds
namespace: registry
spec:
schedule: "*/15 * * * *"
concurrencyPolicy: Forbid
successfulJobsHistoryLimit: 2
failedJobsHistoryLimit: 3
jobTemplate:
spec:
backoffLimit: 1
template:
spec:
restartPolicy: Never
containers:
- name: prune
image: python:3.12-alpine
command: ["python", "/scripts/prune.py"]
volumeMounts:
- name: script
mountPath: /scripts
- name: auth
mountPath: /auth
readOnly: true
volumes:
- name: script
configMap:
name: registry-retention-script
- name: auth
secret:
secretName: registry-regcred

View file

@ -0,0 +1,77 @@
apiVersion: v1
kind: ConfigMap
metadata:
name: registry-retention-script
namespace: registry
data:
prune.py: |
import base64, json, re, urllib.request, urllib.error
REG='https://registry.nxtgauge.com'
CFG='/auth/.dockerconfigjson'
PATTERN=re.compile(r'^[0-9a-f]{40}$')
with open(CFG,'r') as f:
dcfg=json.load(f)
auth=dcfg['auths']['registry.nxtgauge.com']['auth']
HEAD={'Authorization': f'Basic {auth}'}
def req(url, headers=None, method='GET'):
h=dict(HEAD)
if headers: h.update(headers)
r=urllib.request.Request(url, headers=h, method=method)
with urllib.request.urlopen(r, timeout=30) as resp:
return resp.status, dict(resp.headers), resp.read()
_, _, body = req(f'{REG}/v2/_catalog?n=1000')
repos=[r for r in json.loads(body.decode()).get('repositories',[]) if r.startswith('nxtgauge-')]
deleted=0
for repo in sorted(repos):
try:
_, _, tb=req(f'{REG}/v2/{repo}/tags/list')
tags=(json.loads(tb.decode()).get('tags') or [])
except Exception as e:
print(f'[{repo}] tags/list failed: {e}')
continue
sha=[t for t in tags if PATTERN.match(t)]
if len(sha)<=1:
print(f'[{repo}] sha={len(sha)} no prune')
continue
rows=[]
for t in sha:
created='1970-01-01T00:00:00Z'
digest=None
try:
_, h, mb=req(f'{REG}/v2/{repo}/manifests/{t}', headers={'Accept':'application/vnd.docker.distribution.manifest.v2+json'})
digest=h.get('Docker-Content-Digest')
m=json.loads(mb.decode())
cfg=(m.get('config') or {}).get('digest')
if cfg:
_, _, cb=req(f'{REG}/v2/{repo}/blobs/{cfg}')
created=json.loads(cb.decode()).get('created', created)
except Exception:
created='9999-12-31T23:59:59Z'
rows.append((created, t, digest))
rows.sort(key=lambda x: x[0], reverse=True)
KEEP_N=3 # keep last 3 SHA builds (was 1; bumped to prevent auth-blast-radius wipeouts)
keep_set=set(t for _, t, _ in rows[:KEEP_N])
# preserve any -latest aliases regardless of age
keep_set.update(t for t in tags if t.endswith('-latest'))
keep_list=sorted(keep_set)
print(f'[{repo}] sha_total={len(rows)} keep={keep_list} remove={max(0, len(rows)-len(keep_set))}')
for _, t, d in rows:
if t in keep_set or not d:
continue
try:
req(f'{REG}/v2/{repo}/manifests/{d}', method='DELETE')
deleted+=1
print(f' deleted {repo}:{t}')
except urllib.error.HTTPError as e:
print(f' delete failed {repo}:{t} code={e.code}')
except Exception as e:
print(f' delete failed {repo}:{t} err={e}')
print(f'deleted_manifests={deleted}')