Upgrading WarehousePG Enterprise Manager

Upgrade Collector before upgrading WarehousePG Enterprise Manager (WEM). No service shutdown is required before running the upgrade commands.

Before you upgrade

Review the section for the version you're upgrading to before running any upgrade commands. Some of these changes need action on your part beforehand, not just awareness.

Upgrading to 1.3

WEM requires a PostgreSQL database for its own application state, either your WHPG cluster's own database or a dedicated instance. If you don't already have one running, see Installing PostgreSQL.

WEM 1.3 introduces the acp-host-agent service for privileged host operations. See Installing the host agent.

WEM 1.3 bundles Alertmanager and runs its own instance on 127.0.0.1:9093, and the ALERTMANAGER_URL parameter is removed from /etc/wem/wem.conf. If you had a separately installed Alertmanager configured through it, WEM stops using it after the upgrade. If that separate instance is still running on the same host, it might block WEM's bundled instance from starting. Before upgrading, shut it down if it existed only to serve WEM, or move it to a different port first if you still need it.

Before upgrading, check whether you hand-added any custom alert rules to WEM's previous Prometheus rule files (config/prometheus/alerts/*.rules.yml). Alert rules move into WEM's own database in 1.3, so any custom rules from 1.2 are lost once the old rule files are removed. Note down their PromQL expressions and thresholds beforehand, and recreate them through the Rules tab after upgrading. See Creating and managing alert rules.

The RPM preserves your existing wem.conf and places a reference copy of the updated configuration at wem.conf.rpmnew, including new parameters such as WEM_AGENT_CALLBACK_BASE_URL. WEM 1.3 introduces many new and changed parameters beyond this one example. After upgrading, review wem.conf.rpmnew alongside the full Configuration reference and copy any new parameters you need into your existing wem.conf.

Upgrading to 1.2

WEM 1.2 changes the service identity from root to gpadmin. The upgrade handles both common cases automatically:

gpadmin exists on the WEM host (typical when WEM is co-located with WHPG): The upgrade adjusts file ownership so the gpadmin-run service can write logs. The service user change takes effect on systemctl restart wem.
gpadmin is missing (older standalone WEM host that ran as root): The upgrade automatically sets WHPG_ALLOW_ROOT=1 in wem.conf so the service continues running as root, with instructions written to the service configuration for switching to gpadmin later.

If you have a custom service configuration override that sets User=root, the service refuses to start after the upgrade. Either remove the override or set WHPG_ALLOW_ROOT=1 in /etc/wem/wem.conf.

Upgrading Collector

Run these commands from the WarehousePG coordinator.

On the coordinator, download the new Collector package from the EDB repository:

export EDB_SUBSCRIPTION_TOKEN=<your-token>
export EDB_REPO=gpsupp
curl -1sSLf "https://downloads.enterprisedb.com/$EDB_SUBSCRIPTION_TOKEN/$EDB_REPO/setup.rpm.sh" | sudo -E bash
sudo dnf download edb-whpg-observability-collector

Create a file named all_hosts listing every host in the cluster:
```
cdw
scdw
sdw1
sdw2
```

Copy the package to all hosts:

WHPG 7
WHPG 6

gpsync -f all_hosts edb-whpg-observability-collector*.rpm =:/tmp/

gpscp -f all_hosts edb-whpg-observability-collector*.rpm =:/tmp/

Upgrade the package on all hosts:

gpssh -f all_hosts -u gpadmin -e "sudo dnf upgrade -y /tmp/edb-whpg-observability-collector*.rpm"

Redeploy the updated Collector service across all hosts:
```
cd /var/lib/whpg-observability-collector/
./deploy-observability
```

Upgrading WEM

Run these commands on the WEM host:

Configure the EDB repository:

export EDB_SUBSCRIPTION_TOKEN=<your-token>
export EDB_REPO=gpsupp
curl -1sSLf "https://downloads.enterprisedb.com/$EDB_SUBSCRIPTION_TOKEN/$EDB_REPO/setup.rpm.sh" | sudo -E bash

Upgrade the WEM package:
```
sudo dnf upgrade whpg-enterprise-manager
```
See Before you upgrade for version-specific changes that affect this step.
Restart the WEM service to apply the update:
```
systemctl restart wem
```
Clear your browser cache or open WEM in a new browser session. After an upgrade, stale session data can cause a CSRF token error on first login.

WEM 1.2 introduces an optional remote write receiver that allows WEM to accept inbound metrics from Grafana Alloy or any Prometheus-compatible client via a POST /api/v1/write endpoint. This feature is disabled by default and requires explicit configuration to enable. See Remote write receiver (optional) in the Configuring WEM section for details.

Upgrading whpg-enterprise-manager also upgrades its bundled dependencies.

Installing the host agent

The acp-host-agent service is new in WEM 1.3. If you're upgrading from an earlier version, install it on the coordinator and standby coordinator hosts. See Installing the host agent for the full procedure.

Next steps

Your upgrade is complete. See Getting started with WEM.