fix(container): update image kopiur ( 0.10.6 → 0.10.7 )

[renovate]

This PR contains the following updates:

Package	Update	Change
kopiur	patch	0.10.6 → 0.10.7

---

Release Notes

home-operations/kopiur (kopiur)

v0.10.7

Compare Source

Features

• per-repository mover-Job concurrency limits + jitter hardening (#​437) (1529aa9)

---

Configuration

📅 Schedule: (in timezone America/Chicago)

• Branch creation
  • At any time (no schedule defined)
• Automerge
  • At any time (no schedule defined)

🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.

♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

🔕 Ignore: Close this PR and you won't be reminded about this update again.

---

• If you want to rebase/retry this PR, check this box

---

This PR was generated by Mend Renovate. View the repository job log.

[github-actions]

Path: infrastructure/kopiur
Version: 0.10.6 -> 0.10.7

--- /tmp/tmp.FnwqybMqtk	2026-09-02 22:31:03.963559074 +0000
+++ /tmp/tmp.eEDPCCBNUI	2026-09-02 22:31:04.484540071 +0000
@@ -399,7 +399,7 @@
           type: RuntimeDefault
       containers:
         - name: controller
-          image: ghcr.io/home-operations/kopiur-controller@sha256:93f10fd1be0350dac34b80e5a84953548ca7ee3ef04fe36de9179954563bb65e
+          image: ghcr.io/home-operations/kopiur-controller@sha256:ce50917db136e4f19b480724d5cff7a4e1eeed9431e737aeb7808b2693dcefe5
           imagePullPolicy: IfNotPresent
           args:
             - --leader-elect=true
@@ -419,8 +419,8 @@
               value: "release-name-kopiur"
             # The mover image the controller stamps into every Backup/Restore Job.
             - name: KOPIUR_MOVER_IMAGE
-              value: "ghcr.io/home-operations/kopiur-mover@sha256:2c53c0d2f4bf647eb36aebd033f\
-                014e798ec2bb2c08a48bcf3a2dedf21d8a3e7"
+              value: "ghcr.io/home-operations/kopiur-mover@sha256:b6efbb3dedb88a87bc01f089f5f\
+                5e61dc3b306a4003a604c3bf01b1d3c335033"
             - name: KOPIUR_MOVER_PULL_POLICY
               value: "IfNotPresent"
             # The mover PATCHes Backup/Restore .status, so its Job pods run as a
@@ -451,6 +451,13 @@
             # opt-in backstop.
             - name: KOPIUR_MAX_CONCURRENT_DELETE_JOBS
               value: "0"
+            # Cluster-wide cap on concurrently running POOLED mover Jobs
+            # (backups, restores, replication sources) across every
+            # repository — the backstop beneath each repository's own
+            # spec.concurrency.maxConcurrentJobs. 0 = uncapped (default);
+            # restores are always admitted and never queued.
+            - name: KOPIUR_MAX_CONCURRENT_JOBS
+              value: "0"
             # Per-controller cap on concurrent reconciles. Bounds apiserver
             # load and fds during re-list storms/outages; 0 = unbounded (the
             # pre-fix behavior; not recommended).
@@ -545,7 +552,7 @@
           type: RuntimeDefault
       containers:
         - name: webhook
-          image: ghcr.io/home-operations/kopiur-webhook@sha256:b5fdc270d826e329a3a47b5e96a29e412b83332fe5176d1f63c349ffeabb955b
+          image: ghcr.io/home-operations/kopiur-webhook@sha256:788171c81579946053479fb81c7762564471761c3bdf45721dc6a0eacebfc6fa
           imagePullPolicy: IfNotPresent
           env:
             - name: RUST_LOG
@@ -1148,6 +1155,51 @@
               KopiurRepositoryBreakerOpen / the repository's BackendReachable
               condition). They will launch automatically once the repository
               recovers."
+        - alert: KopiurSnapshotWaitingForSlot
+          # A backup queued behind its repository's mover-Job concurrency cap
+          # (spec.concurrency.maxConcurrentJobs, or the cluster-wide
+          # KOPIUR_MAX_CONCURRENT_JOBS backstop) for 30m.
+          #
+          # Deliberately a SEPARATE alert from KopiurSnapshotsGated, on a separate
+          # gauge: a gated Snapshot is waiting on a BROKEN repository, a queued one
+          # is waiting on a WORKING repository that is merely busy. Same symptom
+          # (Pending, no Job), opposite remediation — one is "fix the backend", the
+          # other is "raise the cap or spread the schedules" — so folding them into
+          # one alert would send every reader down the wrong path half the time.
+          #
+          # PER-SERIES, not summed: kopiur_snapshot_waiting_for_slot is one series
+          # per queued CR, and 30m is a long time for ONE run to sit in line — the
+          # thing worth paging on is a queue that is not draining, which a summed
+          # depth cannot distinguish from a healthy queue with fast turnover. A run
+          # that gets its slot removes its own series (the gauge drains to absence,
+          # never to 0), so `for: 30m` means "this specific run never moved".
+          #
+          # max-by for scrape-target-overlap identity hygiene, matching
+          # KopiurRepositoryBreakerOpen; a deposed replica's stale sample is evicted
+          # by scrape staleness (~5m) long before the 30m `for:` completes.
+          #
+          # NOTE `namespace` here is the SNAPSHOT's namespace, not the repository's
+          # — the repository is named by repository_kind/repository.
+          expr: max by (repository_kind, repository, namespace, name)
+            (kopiur_snapshot_waiting_for_slot) == 1
+          for: 30m
+          labels:
+            severity: warning
+          annotations:
+            summary: "Backup {{ $labels.namespace }}/{{ $labels.name }} has been queued for
+              a mover slot for 30m"
+            description: "Snapshot {{ $labels.namespace }}/{{ $labels.name }} has been
+              waiting 30m for a slot in {{ $labels.repository_kind }} {{
+              $labels.repository }}'s mover-Job pool. The repository is working
+              — it is at its concurrency cap — and the run will start on its own
+              when a slot frees; nothing is lost. If this persists, the cap is
+              below what the schedules pointed at that repository need: raise
+              spec.concurrency.maxConcurrentJobs (or the cluster-wide
+              maxConcurrentJobs), spread the schedules with
+              scheduleDefaults.jitter, or check for a wedged mover Job that is
+              never terminating. Beware that while runs queue, a schedule with
+              startingDeadlineSeconds set may be permanently SKIPPING slots
+              (SkipExpiredSlot events)."
 
 ---
 # Source: kopiur/charts/kopiur/templates/servicemonitor.tpl