Managing multiple SolarWinds Orion instances in PocketNOC

Last updated: 2026-05-24

Add more than one Orion server to PocketNOC, name them so the on-call engineer knows which environment they're looking at, and filter alerts per instance.

Overview

Most serious SolarWinds shops run more than one Orion instance — a production Orion managing real infrastructure, plus one or more isolated instances for staging, customer environments, lab work, or geographically separated regions. PocketNOC supports adding multiple Orion connections to the same app and switching between them. This guide covers the setup, naming convention, and the per-instance alert filtering pattern that keeps the on-call experience sane.

When you actually need this

Production + staging that an engineer monitors during deploys.
Multi-region: US Orion, EU Orion, APAC Orion, each managing local infrastructure.
MSP / consulting: one PocketNOC app, several customer Orion environments.
A primary Orion plus a secondary one running parallel for a migration.

If you have one Orion handling everything, skip this guide — the default single-instance flow in the main setup walkthrough is what you want.

Prerequisites

Multiple Orion servers, each reachable from the device (whether via VPN, Tailscale, Cloudflare Tunnel, corporate network, etc.).
A dedicated read-only Orion account on each instance. Per-instance accounts isolate any credential rotation or revocation.
PocketNOC Pro subscription.

Setup

1. Add the first instance (if you haven't already)

Follow the main setup walkthrough. After login, the app stores the connection under whatever name you gave it.

2. Add the second instance

In PocketNOC, open Settings → Orion Servers → Add Server. Enter:

Display name: something unambiguous (see naming convention below).
Host / port: the second Orion server's hostname and SWIS port.
Username / password: the dedicated read-only account on that instance.

On first connect, accept the cert fingerprint via TOFU — exactly as you did for the first instance. PocketNOC stores credentials per instance in the device keychain.

3. Repeat for each additional Orion

There is no hard cap on instance count in PocketNOC. Practical advice: more than 4-5 active connections gets unwieldy on the small screen, and that's usually a sign that the team should standardize on a primary Orion plus an aggregator (Orion's Enterprise Operations Console, or a federated alerting tool) rather than juggling instances from the app.

Naming convention

The display name is what shows up at the top of every dashboard, on every alert push, and in the server-switcher menu. Make it impossible to confuse environments. Some patterns that work:

prod-orion / staging-orion / lab-orion — short, deterministic, no ambiguity.
us-east-orion / eu-west-orion / apac-orion for regional splits.
acme-prod / acme-staging / globex-prod for MSP setups (customer name first).

What to avoid:

orion as the name of any instance when you have more than one. The point of the display name is to disambiguate; "orion" doesn't.
Color or emoji to distinguish. The push notification UI doesn't render them reliably across iOS / Android.
IP addresses or hostnames. prod-orion is meaningful; 10.42.0.1 is not, especially at 3am.

Per-instance alert filtering

Once you have multiple instances, the firehose of alerts gets noisy fast. PocketNOC's alert screen has a server filter at the top — toggle to limit the view to one instance. For push notifications, the per-instance toggle is under Settings → Orion Servers → tap an instance → Push Notifications.

Common pattern:

Production: push enabled for Critical and Warning.
Staging: push enabled for Critical only.
Lab: push disabled. Visit the alerts screen on demand instead.

The result: the phone interrupts the on-call engineer for things that actually warrant interruption, and the noisier non-prod instances are still visible when needed without competing for attention.

What this does NOT do

PocketNOC does not aggregate node lists or dashboards across instances. The Summary Dashboard, Node Health Dashboard, and per-module screens are scoped to one instance at a time — switch instances using the server picker in the top nav. If you need a single pane showing health across all Orion instances, that's what SolarWinds' Enterprise Operations Console (EOC) is for, on the desktop. PocketNOC is intentionally scoped to the per-instance view.

SolarWinds account recommendations (per instance)

Create a dedicated read-only account on each Orion instance. Don't reuse the same credentials across environments — production / staging / lab should each have their own.
Active Directory accounts work if your domain trust spans the environments; otherwise local Orion accounts per instance.
For MSP setups: each customer's Orion gets its own credential. If a customer relationship ends, you revoke one account on one instance and the app loses access to only that customer's data.

Troubleshooting

Pushes from the wrong instance. The per-instance push toggle (Settings → Orion Servers → instance → Push Notifications) is independent per server. Confirm the right instances are enabled.

Confusing which instance you're on. Make the display names obvious (see the naming convention section). PocketNOC also shows the active instance in the top header of every screen.

Can't add a new instance — "connection refused." Same root causes as a single-instance setup: the device can't reach the new Orion's SWIS port. Check the transport (VPN, tunnel, network), then verify the port with a browser visit before retrying in the app.

Different cert fingerprints between instances. Expected — each Orion server has its own cert. PocketNOC pins per-instance, so a change on one server won't trigger warnings on the others.

Security considerations

Per-instance credentials limit blast radius. A compromised phone gives access to every instance you've added, but each instance's account is read-only and scoped — revoke them per-instance if the phone is lost.
Biometric lock + privacy screen apply globally to the app; once unlocked, all instances are accessible. Set a short biometric timeout if you handle multiple customer environments.
For MSP setups, document which engineers have which customer Orion in PocketNOC. Treat the instance list itself as customer-data-adjacent.