* support run prometheus server as statefulset
set server.workloadKind=StatefulSet to run as statefulset
Signed-off-by: Feng Shao <shaof777@gmail.com>
* change per review comment
change boolean flag useStatefulSet in values to controll deployment
type.
bump the chart to a new minor version since a new parameter is added.
Signed-off-by: Feng Shao <shaof777@gmail.com>
* update changelog
Signed-off-by: Feng Shao <shaof777@gmail.com>
* set update strategy and serviceName separately
Signed-off-by: Feng Shao <shaof777@gmail.com>
* fix: dynamically set apiVersion and kind for server deployment
Signed-off-by: Jota Martos <jota.martos@broadcom.com>
---------
Signed-off-by: Feng Shao <shaof777@gmail.com>
Signed-off-by: Jota Martos <jota.martos@broadcom.com>
Co-authored-by: Jota Martos <jota.martos@broadcom.com>
* [bitnami/thanos] per-shard extraFlags + shardLoopId in sharded storegateway
Brings the sharded storegateway in line with capabilities other sharded fields already provide:
- Each `timePartitioning` entry may now set its own `extraFlags` list, appended to the global `storegateway.extraFlags`. Mirrors the existing per-partition `resources` override.
- `topologySpreadConstraints` is now rendered with `shardLoopId` in context, matching what `affinity` / `nodeSelector` / `tolerations` already do. Users can write a per-shard label selector (e.g. `shard: '{{ .shardLoopId }}'`) so each shard's replicas spread independently.
Also defines `$partition` once at the top of the iteration loop rather than re-extracting it via `slice ... | first` in each block.
Signed-off-by: Elliot Kennedy <me@elken.io>
* Update CHANGELOG.md
Signed-off-by: Bitnami Bot <bitnami.bot@broadcom.com>
---------
Signed-off-by: Elliot Kennedy <me@elken.io>
Signed-off-by: Bitnami Bot <bitnami.bot@broadcom.com>
Co-authored-by: Bitnami Bot <bitnami.bot@broadcom.com>
* [bitnami/mongodb-sharded] fix: use exact match in liveness probe pgrep
The mongos liveness probe uses `pgrep mongos` to check if the mongos
process is running. However, `pgrep` performs substring matching by
default, so `pgrep mongos` also matches `mongosh` processes (since
"mongos" is a substring of "mongosh").
This causes a critical failure mode during startup: when all MongoDB
sharded StatefulSets start simultaneously, mongos may attempt to connect
to configsvr before its replica set is fully initialized. The Bitnami
entrypoint script runs `mongosh --host configsvr` to verify configsvr
availability, but this call blocks indefinitely when configsvr's port is
open but the replica set hasn't completed primary election or auth user
creation. Since the `mongosh` call has no timeout, it hangs forever,
and the actual `mongos` router process is never started.
At this point, the liveness probe should detect that mongos is not
running and restart the container. However, `pgrep mongos` matches the
hung `mongosh` child process, so the liveness probe passes. The
container remains stuck permanently: liveness passes (due to the
substring match on mongosh), readiness fails (correctly, since mongos
never started), and Kubernetes never restarts it.
Meanwhile, shard data nodes complete their own init, stop mongod, and
try to register with mongos. Since mongos never started, shards loop on
"timeout reached before the port went into state inuse" and get killed
by their own liveness probes, entering CrashLoopBackOff.
The fix adds the `-x` flag to `pgrep`, which requires an exact match on
the process name. With this change, `pgrep -x mongos` matches only the
real `mongos` process and not `mongosh`, so if the entrypoint hangs, the
liveness probe correctly fails and Kubernetes restarts the container.
On retry, configsvr is typically ready, and startup succeeds.
The same `-x` fix is applied to the other liveness probes in this chart
that suffer from the same substring-match weakness: the config-server
and shard data StatefulSets (both using `pgrep mongod`, which can match
processes such as `mongodump` or any other `mongod*` binary). The shard
arbiter probe is also switched from `pgrep -f mongodb` to `pgrep -x
mongod` so that all mongod-based StatefulSets check for the actual
process by exact name rather than relying on substring or full-cmdline
matches.
Signed-off-by: delthas <delthas@dille.cc>
* Update CHANGELOG.md
Signed-off-by: Bitnami Bot <bitnami.bot@broadcom.com>
---------
Signed-off-by: delthas <delthas@dille.cc>
Signed-off-by: Bitnami Bot <bitnami.bot@broadcom.com>
Co-authored-by: Bitnami Bot <bitnami.bot@broadcom.com>
* feat(nginx): add checksum annotation for existingServerBlockConfigmap
When existingServerBlockConfigmap or existingStreamServerBlockConfigmap
are set, use helm lookup to read the ConfigMap from the cluster and
compute a checksum annotation on the pod template. This triggers a
rolling restart whenever the referenced ConfigMap content changes,
consistent with the behavior already provided for inline serverBlock.
Also fix a copy-paste error: the stream server block checksum condition
was checking existingServerBlockConfigmap instead of
existingStreamServerBlockConfigmap.
Note: lookup returns nil during helm template (offline dry-run), so
the annotation is silently omitted in that case. This is expected
and consistent with how lookup behaves across all Helm charts.
Signed-off-by: Andrea Pilo <apilo@cuebiq.com>
* Update CHANGELOG.md
Signed-off-by: Bitnami Bot <bitnami.bot@broadcom.com>
---------
Signed-off-by: Andrea Pilo <apilo@cuebiq.com>
Signed-off-by: Bitnami Bot <bitnami.bot@broadcom.com>
Co-authored-by: Bitnami Bot <bitnami.bot@broadcom.com>
* feat(haproxy): add IPv6 dual-stack support with ipFamilyPolicy and ipFamilies
Signed-off-by: Awan <jesse.awan@sap.com>
* fix(haproxy): bump chart version to 3.0.1
- Update Chart.yaml version from 3.0.0 to 3.0.1
- Update CHANGELOG.md version number to match Chart.yaml
This fixes CI validation that requires version bump when chart files are modified.
Signed-off-by: Awan <jesse.awan@sap.com>
* Update CHANGELOG.md
Signed-off-by: Bitnami Bot <bitnami.bot@broadcom.com>
* Update CHANGELOG.md
Signed-off-by: Bitnami Bot <bitnami.bot@broadcom.com>
---------
Signed-off-by: Awan <jesse.awan@sap.com>
Signed-off-by: Bitnami Bot <bitnami.bot@broadcom.com>
Co-authored-by: Bitnami Bot <bitnami.bot@broadcom.com>
* [bitnami/keycloak] Fix health probe paths with httpRelativePath
Startup, liveness, and readiness probe paths were hardcoded
without accounting for the `httpRelativePath` value, causing
probes to fail when a custom relative path is configured.
Prefix each health endpoint (`/health/started`, `/health/live`,
`/health/ready`) with the configured `httpRelativePath`, trimming
any trailing slash to avoid double slashes in the path.
Signed-off-by: Robbie Blaine <robbieblaine.rb@gmail.com>
* Update CHANGELOG.md
Signed-off-by: Bitnami Bot <bitnami.bot@broadcom.com>
---------
Signed-off-by: Robbie Blaine <robbieblaine.rb@gmail.com>
Signed-off-by: Bitnami Bot <bitnami.bot@broadcom.com>
Co-authored-by: Bitnami Bot <bitnami.bot@broadcom.com>
* fix: correct context reference for custom labels in VerticalPodAutoscaler metadata
Fixes#36439
fix: update apiVersion for Vertical Pod Autoscaler to use autoscaling.k8s.io/v1
Fixes#36303
chore: bump ClickHouse chart version to 9.4.6
chore: bump version to 2.31.8 in Chart.yaml
Signed-off-by: Per Osbäck <per@osbeck.com>
* Apply suggestion from @jotamartos
Signed-off-by: Juan José Martos <jota.martos@broadcom.com>
* Remove old VPA apiVersion reference
Removed outdated Vertical Pod Autoscaler apiVersion handling.
Signed-off-by: Juan José Martos <jota.martos@broadcom.com>
---------
Signed-off-by: Per Osbäck <per@osbeck.com>
Signed-off-by: Juan José Martos <jota.martos@broadcom.com>
Co-authored-by: Juan José Martos <jota.martos@broadcom.com>
