K-C11, closes#11. On top of #27 (merge that first). Needs server 27.1-rc.34 (inbuxa/inbuxa-server#491): hold until it ships.
server.backup: a nightly CronJob, inbuxa --export s3://..., streamed while the server runs. Keys from existingSecret as files, or the pod's role (serviceAccountName). Old exports are left to a bucket lifecycle rule.
server.storeCheck: quick nightly, full weekly. A check is never retried: exit 3 is an answer.
With the metrics rules and kube-state-metrics: InbuxaBackupMissing (no finished export for backupHours) and InbuxaStoreCheckFailed.
The jobs NetworkPolicy covers both.
Tested: tools/check.sh (93 unit tests). On kind: inbuxa/inbuxa-k8s-lab backup-test.sh runs the backup CronJob, restores into fresh stores, and runs both check CronJobs: all pass. The two alerts weren't run (no kube-state-metrics in the lab).
K-C11, closes #11. On top of #27 (merge that first). Needs server 27.1-rc.34 (inbuxa/inbuxa-server#491): hold until it ships.
- `server.backup`: a nightly CronJob, `inbuxa --export s3://...`, streamed while the server runs. Keys from `existingSecret` as files, or the pod's role (`serviceAccountName`). Old exports are left to a bucket lifecycle rule.
- `server.storeCheck`: quick nightly, full weekly. A check is never retried: exit 3 is an answer.
- With the metrics rules and kube-state-metrics: InbuxaBackupMissing (no finished export for `backupHours`) and InbuxaStoreCheckFailed.
- The jobs NetworkPolicy covers both.
**Tested**: `tools/check.sh` (93 unit tests). On kind: inbuxa/inbuxa-k8s-lab `backup-test.sh` runs the backup CronJob, restores into fresh stores, and runs both check CronJobs: all pass. The two alerts weren't run (no kube-state-metrics in the lab).
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
K-C11, closes #11. On top of #27 (merge that first). Needs server 27.1-rc.34 (inbuxa/inbuxa-server#491): hold until it ships.
server.backup: a nightly CronJob,inbuxa --export s3://..., streamed while the server runs. Keys fromexistingSecretas files, or the pod's role (serviceAccountName). Old exports are left to a bucket lifecycle rule.server.storeCheck: quick nightly, full weekly. A check is never retried: exit 3 is an answer.backupHours) and InbuxaStoreCheckFailed.Tested:
tools/check.sh(93 unit tests). On kind: inbuxa/inbuxa-k8s-labbackup-test.shruns the backup CronJob, restores into fresh stores, and runs both check CronJobs: all pass. The two alerts weren't run (no kube-state-metrics in the lab).432208a7fcto1fce1c4e7b