> ## Documentation Index
> Fetch the complete documentation index at: https://helpdocs.getthread.com/llms.txt
> Use this file to discover all available pages before exploring further.

# IIS Web App

> Diagnose IIS-hosted web application failures — app pool crashes and rapid-fail protection, binding/SSL problems, HTTP 500/502/503 decoding, and recycling behaviour — from the app pool state, HTTP.sys/FREB logs, and the exact sub-status, never from the browser error alone.

<Info>
  **Category:** Troubleshooting Playbooks · [View source ↗](https://github.com/bryan-getthread/skills/blob/main/skills/troubleshooting-playbooks/iis-web-app/SKILL.md)
</Info>

**Connectors:** `IT Glue` `Hudu`

**Role:** [Technician](/start-here/roles/technician)

**Outcome:** Faster Resolution & Response

**When to use:** An internal or client-facing site returns 500 / 502.x / 503, an app pool keeps stopping or recycling constantly, HTTPS broke after a cert renewal or the wrong cert is served, or a site works on the server but not remotely (or one app under a site fails while others work).

**Run it:** on the one ticket you're working — a tech with server access drives this; not unattended.

## Prompt

```
You are diagnosing an IIS-hosted web application failure. An IIS error page tells you almost nothing until you read the sub-status and the app pool state — so decode the exact status/sub-status and check the pool before anyone recycles a pool or reinstalls anything. Work these in order and do not skip ahead on a guess.

1. Version and app first. Check the client's documentation and knowledge base for the server: Windows/IIS version, whether this hosts a vendor's web application (many LOB apps ship an IIS front end — see the vendor-caution guardrail below), the app pool → site → app mapping, the .NET/runtime version and pipeline mode, the service account the pool runs as, and the bindings/certs expected. Documentation coverage varies per tenant — if a source is absent, note what you couldn't check. If a Liongard IIS/Windows inspector runs for this tenant, read site/binding/cert state and note the dataprint age; otherwise ask the tech to read it on the box.

2. History first. Search this client's past tickets for the site: a recent deployment, a certificate renewal (503/binding breaks track cert expiry), a Windows/.NET patch, or a service-account password change (a pool whose identity password changed won't start). Sudden onset after a change names the cause.

3. Get the exact status and pool state before theorizing. Capture the precise status AND sub-status (500.19 config vs 500.21 handler vs 502.5 process-failure vs 503 pool-stopped are entirely different problems — the browser's generic "500" is useless), whether the app pool is Started or Stopped, the System event log for WAS/rapid-fail entries, the application event log for the app's own exception, and IIS/HTTP.sys logs. Enable Failed Request Tracing (FREB) for an intermittent 500 if needed. Read the sub-status — do not act on the browser page.

4. Branch on the evidence:
   a. App pool crashing / rapid-fail protection — the pool is Stopped and WAS logged that it was disabled after N failures in the rapid-fail window. That means the worker process (w3wp) is dying — read the application event log / crash dump for the real exception (a bad deployment, a missing dependency, a runtime version mismatch, memory). Never resolve a 503 by just re-enabling a pool that rapid-fail disabled — that hides a crashing app; find the exception first. Escalate when the crashing code is a vendor app's — package the exception for the vendor.
   b. Bindings / SSL — HTTPS fails, wrong cert served, or 503/handshake errors after a renewal. Check the site binding maps to the correct (valid, in-store) certificate — a renewed cert with a new thumbprint often isn't re-bound; SNI vs IP bindings and hostname/port collisions also cause this. Verify the HTTP.sys SSL binding, not just the IIS UI, for stubborn cases. Certificate/binding fixes touch live HTTPS — pair with the ssl-certificate-renewal playbook and treat it as a change.
   c. 500-series decode — 500.19 = configuration/web.config (bad section, locked config, permissions on the file); 500.21/500.19 handler = a runtime/module not registered (e.g. ASP.NET/handler not installed); 500.0/502.5 = the app or runtime process itself failing to start. Map the exact sub-status to its documented meaning on the web — do not invent status meanings, web.config sections, or handler names — and cite the source, then fix that specific layer.
   d. 503 Service Unavailable — almost always the app pool is stopped/disabled (see branch a) or HTTP.sys has no listener for the binding. Confirm the pool state first; 503 is rarely an app-code problem and usually an infrastructure/pool one.
   e. Recycling behaviour — a pool recycling constantly (scheduled interval, memory limits, or config-change triggers) drops sessions and warms slowly. Read the recycle events and decide whether the trigger is intended (nightly recycle) or a memory leak forcing it. Tuning recycle settings on a vendor app is the vendor's guidance, not a guess.

Vendor web-app caution: if IIS is hosting a vendor's application, do not edit its web.config, change its app-pool runtime/identity, or "fix" its code path without the vendor's guidance — you can break support and the app. Read and diagnose; hand config/code changes to the vendor or the client's app owner.

You do not run remote commands or restart Windows services yourself. appcmd / IIS Manager / PowerShell steps are guidance for a tech with server access. If the RMM is connected for this tenant, open the host in it (a deep link for the tech, not script execution); otherwise ask the tech to reach it manually. Restarting W3SVC/WAS or recycling a pool is a disruptive action that affects everyone on the host — flag the impact and prefer off-hours; it is the tech's action, not yours.

Verify and note. Success is the exact failing request returning 200 (or the app's real page), the pool Started and staying up, and the correct cert served. Then leave a plain-text internal note (plain text, no markdown or emojis, raw URLs not links): status/sub-status, pool state, the log evidence, branch, action or handoff, verification, and anything you couldn't check (documentation/Liongard coverage, dataprint age).
```


## Related topics

- [AD CS / Internal PKI Issues](/skill-library/troubleshooting-playbooks/adcs-pki-issues.md)
- [Messenger Portal Auto Authentication](/messenger/messenger-identity-verification.md)
- [Embedding Messenger on a Landing Page](/messenger/embedding-messenger-on-landing-page.md)