Microsoft 365 services timing out

[munrobasher]

This is a bit of a background kind of question. I've got a VM 200Mbps connection at home and it's rock solid and fast. Zero complaints. I've worked from home as IT specialist for six years. Everything fine except recently I'm getting performance issues with Microsoft 365 services, specifically the various admin consoles. Not just for one tenant/domain either - I support several companies and it's occurring with them all. 

Now of course my initial thoughts is that the M365 servers are overloaded. After all, I've occasionally had similar performance problems (pages not rendering) when working at my main clients site where the internet is via M247 and 1Gbps.

But Microsoft are denying any problems and the fact that if I remote into a laptop at this client's site, the same web pages open fine, it does suggest that it's not a problem with their server.

So what else could this be? My internet connection is fine, every other web site and service (like Outlook/Exchange) works flawlessly. Just get errors like this:

Loading chunk 6330 failed. (error: https://prod.msocdn.com/M365/admin-main/2022.4.14.7/inline.en.6330.chunk.js) df32c1e2-9477-4574-af1d-b160789e9f72

My thought is that Virgin Media have had routing problems in the past and when only one service is effected, one's thoughts turned to invisible downstream proxy/caching servers. Not the first time that invisible caching has caused problems with ISPs.

Next time I get problems, I'm going to connect my laptop to my mobile hotspot and try that route.

[Ryan_N]

Just to be clear - when it does happen, is it often around the same points of the day? Is there any difference any other parts of the day etc? 

Cheers, 

Ryan. 

[munrobasher]

Afraid I haven't been keeping that accurate logs. Whilst I *tend* to work normal-ish hours, I can often do some work in the evening. I've done a bit of M365 admin today and no problems. Which is probably further evidence that it's not router related. If there is a proxy caching server downstream that's overloaded, then this could be variable.

It would be useful if we knew for certain *if* Virgin Media do implement invisible HTTP caching servers.

[Jonny-M]

Out of interest, if you're an M365 admin and you have turned on the network connectivity insights (https://admin.microsoft.com/#/networkperformance) do you ever see an entry for "Intermediary devices are affecting your network"?

The MS docs entry for the error is https://docs.microsoft.com/en-GB/Microsoft-365/Enterprise/office-365-network-mac-perf-insights?WT.mc... 

[munrobasher]

Will enable this later. Too busy at moment but right now, M365 admin pages are timing out (9:34am Friday)

[munrobasher]

Horribly slow this morning. M365 admin consoles timing out on multiple tenants. Every other website is fine. Refresh sometimes gets it working.

[Travis_M]

Hi @munrobasher

Thanks for your reply

Is this even after trying the above suggestion from Jonny-M? How has this been for the last couple of days?

Regards

[IPFreely]

You're connecting to 365 over HTTPS. VM do not and cannot cache HTTPS. They'd need to have a wildcard certificate trusted by your devices and conduct MITM. This would be illegal and technically unfeasible.

Unsure of what the issue is likely to be, don't have any data to base a decision on. Strange routing to some services is possible, VM's transit and peering arrangements are quite bizarre.

An MTU discrepancy breaking some HTTPS sessions is possible and can be due to a few things.

No idea beyond that it's not caching. 

[munrobasher]

Good point re HTTPS and encryption. So back to the drawing board. Completely unusable this morning. I'm now, sadly, contemplating having to leave VM after 25 years as it's critical. My neighbour is on TalkTalk (!) and has let me have their Wi-Fi password. Every time I've had problems, I've switched to their network and the pages load instantly.

There was a really weird period a few months ago when it looked like a failed update reached my VM router. I saw, for a few hours, a new logon page with a picture. I suspect that update was rapidly backed out locally as our local FB lit up with "problems with Virgin Media". Magically, it went back to the old version. It totally trashed my port forwarding rules.

So I'm going to try a factory reset of the router. My next plan is to put it into modem mode and use a spare Mikrotik router I've got lying around. But that's a lot of hassle!

What I haven't tried yet is a different browser from Edge/Chrome which are basically the same rendering engines. I'll give Firefox a go...