GET vs HEAD for traffic manager health probe

[mikewo]

The documentation all indicates that the health probe from Traffic Manager uses a GET request to the URL provided; however, in our systems we are seeing a HEAD call to the URL provided rather than a GET. Is the documentation just incorrect?

---

Document Details

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

• ID: 9fed131c-15d3-3266-3570-08e445eda1b8
• Version Independent ID: fe645535-673e-4c59-0163-be55107b0e58
• Content: Azure Traffic Manager endpoint monitoring
• Content Source: articles/traffic-manager/traffic-manager-monitoring.md
• Service: traffic-manager
• GitHub Login: @asudbring
• Microsoft Alias: allensu

[mikewo]

This is also referenced in the FAQ as using a GET for the health probe at https://docs.microsoft.com/en-us/azure/traffic-manager/traffic-manager-faqs#what-application-protocols-can-i-use-with-traffic-manager.

[msrini-MSFT]

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

[msrini-MSFT]

@mikewo , I did a lab repro and I am able to see only GET request from Azure Traffic Manager.

Frame: Number = 8384, Captured Frame Length = 214, MediaType = ETHERNET

• Ethernet: Etype = Internet IP (IPv4),DestinationAddress:[00-0D-3A-8A-7E-32],SourceAddress:[12-34-56-78-9A-BC]
• Ipv4: Src = 52.240.151.125, Dest = 172.16.14.4, Next Protocol = TCP, Packet ID = 8763, Total IP Length = 200
• Tcp: Flags=...AP..., SrcPort=44928, DstPort=HTTP Alternate(8080), PayloadLen=160, Seq=4211163617 - 4211163777, Ack=2383194767, Win=1029 (scale factor 0x8) = 263424

• Http: Request, GET /
  Command: GET
  • URI: /
    ProtocolVersion: HTTP/1.1
    UserAgent: Azure Traffic Manager Endpoint Monitor
    Accept: /
    accept-encoding: *
    Host: XXXX.trafficmanager.net:8080
    Connection: Close
    HeaderEnd: CRLF

Can you check the User Agent header in your environment where you are receiving HEAD request ?

[mikewo]

Thanks for the pointer. You are correct and I had made an assumption that these calls were coming from TM because it's the only thing pointed at our health endpoint and the sheer number of them seemed to make sense. The user agent isn't TM, and we are looking at where the traffic is coming from. Since we were seeing this same behavior across multiple of our sites we came to the wrong conclusion.

Closing.