Veleis 1.8.8

Added Advanced DNS Monitoring with A, AAAA, CNAME, MX, NS, TXT, SRV, CAA,
and PTR records; typed expectations; resolver and optional authoritative
comparison; resolver-dependent DNSSEC state; semantic change history; and
shared incident/alert integration.
Added revisioned Notification Escalation Policies with ordered absolute-delay
steps, multiple channels, immutable incident snapshots, stop-on-acknowledge,
exactly-once scheduling, cancellation, delivery provenance, active progress,
and a responsive timeline editor.
Added in-place notification-channel editing and safe optional credential
rotation while preserving channel identity, history, type, encryption, and
write-only secret behavior. Channel management is now a compact responsive
list.
Unified Profile, Security, Preferences, Data Retention, and Access below
canonical /settings/* routes with responsive top navigation. The sidebar
account footer now provides a direct neutral Sign Out action and Settings is
the sole general configuration entry.

Schema advances from 37 to 40. Agent metrics use one-day chunks, drop a
redundant large index, and enter Timescale columnstore after one day while
preserving the existing 14-day agent-metric lifecycle.
Fresh installations size PostgreSQL and TimescaleDB from the minimum of host,
current cgroup-v2/cgroup-v1 ancestors, and a validated optional override.
Profiles bound buffers, per-operation memory, connections, and workers. A
32 GiB host constrained to 2 GiB now receives the 2 GiB profile rather than
host-sized settings.

Run sudo veleis backup before upgrading, then sudo veleis upgrade 1.8.8.
The lifecycle creates and verifies another mandatory pre-upgrade backup,
validates the immutable image digest, applies migrations 38–40, and waits for
schema 40 and HTTPS readiness.
Fresh 1.8.8 installations use cgroup-aware managed PostgreSQL tuning.
Historical and administrator-customized settings are never silently
overwritten during upgrade.
Existing installations should first refresh the current lifecycle tooling as
documented in Upgrading, then run:
```
sudo veleis postgres-memory status
```
HEALTHY, WARNING, or UNSAFE is reported with effective memory,
ownership, and live PostgreSQL settings. An unsafe historical installation
can explicitly adopt the Veleis-managed profile with:
```
sudo veleis postgres-memory adopt-managed
```
Adoption backs up both .env and compose.yaml, atomically installs the
managed profile/topology, recreates and health-checks the stack, and restores
both files if validation fails. It restarts PostgreSQL and the application;
schedule a maintenance window. No database data is deleted. Operators who
intentionally maintain custom PostgreSQL configuration remain in control and
may leave it unchanged after reviewing the warning.
Never use swapoff or hand-edit PostgreSQL files as the Veleis remediation
procedure.

Supported upgrade sources: Veleis 1.7.1 and 1.8.0 through 1.8.7. Backup
format remains 1. Recommended Ravyr remains signed 1.8.2; minimum Ravyr is
1.7.0 and lifecycle protocol remains 1.
The supported image remains linux/amd64. Default HTTPS remains self-signed.
Image signing, a published SBOM, and provenance attestation are not currently
part of the Veleis release pipeline.
PostgreSQL memory profiles bound the primary allocation risks but do not
guarantee that every third-party workload fits the minimum 1 GiB profile;
2 GiB is the practical small-install recommendation.
Docker image: docker.io/nyxmael/veleis:1.8.8
Manifest digest: sha256:4f082699b5bec6261f119ea6f620decbf9bcfc72c06523b1e8ae1c8e00a98f8a

Navigation Menu