Ignition Designer Alarm Pipelines Not Opening: NPE Fix Path

Symptom Profile on 8.1.45

The reported environment is Ignition v8.1.45 (b2025010709) with the Designer running on Azul Systems, Inc. 17.0.13. The failure has these parts:

• Clicking Alarm Notification Pipelines in the Project Browser leaves the center workspace on whatever was showing before. The Pipeline Block Editor dockable pane appears, but the pipeline canvas does not.
• After you create a new pipeline, clicking or double-clicking it does nothing. The main window does not respond.
• Perspective and Sequential Function Charts switch to their workspaces normally.
• Alarm Notification Pipelines, Scripting, Named Queries and Web Dev show no workspace when clicked at the top level.
• The Designer Log Viewer records nothing when the clicks fail.
• View > Reset Panels throws a java.lang.NullPointerException.
• The failure reproduces in a new, blank project.
• Deleting the cached runtimes does not clear it.

A top-level folder without a splash screen is not a fault in itself, because splash behavior differs between modules. The conclusive symptom is a resource that will not open. A pipeline you just created that ignores a double-click means the Designer cannot switch to that resource's editor.

Reading the Reset Panels Stack Trace

The Reset Panels exception is the only hard evidence the Designer produces here. Treat it as the primary diagnostic, not as a separate problem:

java.lang.NullPointerException: Cannot invoke "com.inductiveautomation.ignition.designer.model.ResourceWorkspace.getFrames()"
  because the return value of "java.util.List.get(int)" is null
  at com.inductiveautomation.ignition.designer.WorkspaceManager.makeNonSelectedUnavailable(WorkspaceManager.java:214)
  at com.inductiveautomation.ignition.designer.WorkspaceManager.resetToDefault(WorkspaceManager.java:200)
  at com.inductiveautomation.ignition.designer.IgnitionDesigner.resetLayout(IgnitionDesigner.java:727)

The trace shows how the Designer failed:

1. IgnitionDesigner.resetLayout asks the WorkspaceManager to return every workspace to its default docking layout (resetToDefault).
2. makeNonSelectedUnavailable walks the list of registered workspaces. For each workspace that is not currently selected, it calls getFrames() to hide that workspace's dockable frames.
3. One index in that list returns null instead of a ResourceWorkspace, so the call to getFrames() dereferences null.

The workspace manager is holding a slot for a workspace that was never instantiated or was lost. The same manager handles the switch when you click a Project Browser node. If the target workspace is missing, or the switch logic trips over the same null entry, the selection change never completes and the center pane stays where it was. This fits the observed pattern: the module's dockable pane (Pipeline Block Editor) registers, but the workspace that owns the canvas never takes focus.

The affected set mixes platform features (Scripting, Named Queries) with module-provided features (Alarm Notification, Web Dev). A single missing or faulted module does not explain a failure across both groups. The Alarm Notification Pipelines node appears in the Project Browser, which confirms the module loaded into the Designer. The fault sits in the Designer session's workspace state, not in one module's install.

Symptoms Versus Likely Causes

Observation	What it indicates	Next check
Top-level folder click shows no splash	Not diagnostic by itself; splash behavior varies by module	Create a resource of that type and double-click it
Newly created pipeline ignores double-click	Workspace switch to the pipeline editor fails	Run View > Reset Panels and capture the trace
Pipeline Block Editor pane appears, canvas does not	Module Designer hook loaded; workspace selection failed	Compare against the same project from another Designer
Log Viewer is empty on click	Exception is swallowed on the UI thread or never raised on that path	Use Reset Panels to force a logged exception
NPE in makeNonSelectedUnavailable	Workspace list holds a null entry	Escalate with replication steps
Same behavior in a new blank project	Project resources are not the cause	Isolate the Designer client from the Gateway
Persists after deleting cached runtimes	Not a stale-cache fault	Record this for the support case

Isolation Approaches Compared

Every approach below was either tried against this fault or is the standard next step. Judge each one by whether it changes the outcome and whether it tells you where the fault lives.

Approach	Result here	What it rules out	Cost
Delete Designer cache / cached runtimes	No change	Stale cached resources and a bad downloaded runtime	Minutes
Test in a new, blank project	Same failure	Corrupt project resources and project inheritance	Minutes
View > Reset Panels	Throws NPE	Nothing; converts a silent failure into a logged exception	Seconds
Launch a different Designer instance	Symptom not seen	Gateway-side module state, if the same Gateway and project are used	Minutes to an hour
Open a case with Inductive Automation Support	Recommended next step; not a known issue at time of report	Product defect versus local environment	Case turnaround

Cache deletion fixes a large share of short-term Designer faults, so it should come first. Once it fails, repeating it adds nothing. A result that survives a cache wipe and a blank project is useful information for Support, because it removes the two most common explanations.

Recommended Path: Isolate the Designer Client, Then Escalate

Before anything else, confirm whether the fault follows the workstation or the Gateway. That single split decides every later step:

• Fault follows the workstation: the same Gateway and project open correctly from another machine or another OS user profile. Rebuild the local Designer environment.
• Fault follows the Gateway: every Designer connected to that Gateway shows the dead nodes. Check module state on the Gateway and open a support case with Gateway diagnostics attached.
• Fault is intermittent in one session: it appears in one launch and not the next. Capture the trace from a failing session before restarting, because a restart destroys the evidence.

In this case, moving to a different Designer cleared the symptom. That points toward client-side state. Confirm it against the same Gateway and the same project, not only a different Gateway, before you treat it as settled.

Procedure: Client Isolation and Rebuild

1. Capture a baseline. Record the Gateway version and build (8.1.45 (b2025010709) here) and the Designer's Java runtime (Azul 17.0.13 here) from the Designer's About information. Do not move on until both are written down.
2. Force the exception. With a project open, run View > Reset Panels. Copy the full stack trace, including the WorkspaceManager.java line numbers. A second NPE at makeNonSelectedUnavailable confirms the fault is still present in this session.
3. Build a minimal reproduction. Create a new, blank project. Add one alarm notification pipeline, one named query and one project library script. Double-click each one and note which open an editor and which do not.
4. Check module state on the Gateway. On the Gateway web interface, open the module list. Confirm the Alarm Notification and Web Dev modules show as installed and running with no fault. A faulted module here moves the investigation to the Gateway.
5. Launch from a second workstation or OS user. Use a clean Designer Launcher against the same Gateway and open the same test project. Double-click the same three resources. If they open, the fault is local to the first workstation.
6. Rebuild the local Designer environment. On the failing workstation, close all Designer and launcher processes. Clear the launcher cache and cached runtimes under the user profile's .ignition directory, then reinstall the Designer Launcher from the Gateway's download page. Relaunch and repeat step 3.
7. Escalate if the rebuild does not clear it. A reproducible NPE with replication steps is enough to open a case with Inductive Automation Support. Go to the Support Package step below.

If you also run a clean OS user profile test, run it before the full rebuild. It isolates per-user Designer state without touching the working install.

Support Case Package

At the time of the report, this behavior was not listed as a known issue, and it differs from other Designer faults that a cache clear resolves. Give Support a package that removes the first round of questions:

Item	Content
Versions	Gateway v8.1.45 (b2025010709); Designer Java Azul Systems, Inc. 17.0.13; OS of the Designer workstation
Exception	Full java.lang.NullPointerException trace from View > Reset Panels
Replication steps	New blank project, create pipeline, double-click, no editor opens; list of affected nodes (Alarm Notification Pipelines, Scripting, Named Queries, Web Dev)
Contrast	Perspective and Sequential Function Charts open normally
Actions already taken	Cached runtimes deleted with no change; blank project reproduces; result of the second-workstation test
Logs	Designer console output from a failing session and the Gateway wrapper log covering the Designer launch time

State clearly that the fault survived a cache deletion. That result tells Support the issue is not stale cached state and moves the case directly to workspace initialization.

Verification

1. Relaunch the Designer, open the blank test project and run View > Reset Panels. The layout must reset with no exception dialog and no new NPE in the console.
2. Double-click the test pipeline. The pipeline canvas must take over the center workspace, with the Pipeline Block Editor pane docked beside it.
3. Drag one block onto the canvas and save the project. The save must complete without errors.
4. Double-click the test named query, then the project library script. Each must open in its own editor tab.
5. Switch back to Perspective, then to the pipeline, several times. Each switch must change the center workspace on the first click.
6. Close the Designer, relaunch it and open the production project. Repeat steps 2 and 4 on real resources. Close the case, or confirm the rebuild worked, only after this second launch behaves the same as the first.

FAQ