Where to find why health probes are failing?

[oskarm93]

I had health probes failing on one of my instances. I could not figure out why Front Door did not connect to it.
Are there any logs or diagnostics information about the result of health probes and why Front Door might think the back end is down?
I looked in Log Analytics, but only found logs for incoming requests.

---

Document Details

⚠ Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.

• ID: f35a0b39-ab2f-ba8a-0907-e0f3a6b69135
• Version Independent ID: 0e6c8267-a911-9c6b-0522-3ec96753b3ef
• Content: Azure Front Door Service - backend health monitoring
• Content Source: articles/frontdoor/front-door-health-probes.md
• Service: frontdoor
• GitHub Login: @sharad4u
• Microsoft Alias: sharadag

[msrini-MSFT]

@xenalite , Thank you for your feedback. We will investigate and get back to you shortly.

[msrini-MSFT]

@xenalite , What is the probe config that you have defined in AFD?

Can you do a curl from your machine to http/https://backendpool/ and share me the results?

[oskarm93]

I no longer have the health probes failing because I investigated it and found a misconfiguration.
However, there are a number of issues that could be preventing back end health probes from being successful:

1. Wrong host header
2. Wrong host name
3. Unreachable host - may be reachable from my machine but not Front Door
4. Wrong protocol - HTTP / HTTPS probes
5. Wrong health probe endpoint - 404
6. HTTPS health probe returning with self-signed / untrusted certificate

How could I find out what Front Door thinks is wrong?

Here's my config:

[msrini-MSFT]

@xenalite , It is a valid feedback. In Application Gateway we do have an option to see the error description when the health probe fails.

It is a product ask and I will pass on the feedback to the product group. Thank you for your valuable feedback again!!

[StevenB23]

I'd like to request this feature. What would be the best way to get an alert of health probe failure.

[eliagrady]

Thanks @xenalite for your reply - mine was that pesky HTTPS health probe returning with self-signed / untrusted certificate

[joshuadmatthews]

@eliagrady I suspect I am running up on same issue. Seems like a strange requirement with private link being a backend option normally you wouldn't have a signed cert for stuff inside of a private network, but I guess I do after reading this thread.

[akakira]

We have front door classic, and when we have a single backend in the pool, the health checks return good no matter which endpoint we use, when we add a second endpoint the checks fall to 0. I would love to see why this is happening.