Information, Support and FAQ forums for OutagesIOOur FREE software monitors your Internet service and provider then alerts you when you have connection issues or downtime. For home, small business, IT companies and enterprise. Monitor Internet outages, track network connection reliability, and get proof you can use. For Windows 7, 8, 10, Centos 7, 8, Debian 7, Ubuntu, ARM (Raspberry, Tinker Board, etc).
Learn more about www.outagesio.com by visiting our web siteNOTICE: If you previously registered on forums.outages.io, simply use the 'Forgot Password' option to recover your account.
Increased number of Internal disconnects
Hi all. This is a question regarding issues with internet disconnects classed as "Internal". I've looked for other such issues and can't find anything close to the problem I'm having. Bit of background to wade through first.
I have a Basic Phone line DSL connection that it connected to a combined mode/Router (TP-Link Archer VR900). My max speed is 7Mbs. I have Cable connections to my main computers (including the one the Outages is installed on). I have a mesh system elsewhere in the house due to the thickness of the walls (a single wireless point just won't work).
My broadband disconnects on a fairly frequent basis. Either when I receive a phone call (not every call) and usually re-sets very quickly. Open reach have been out and believe it is the wire connection into the house getting corroded (looked fine last time I looked) but I have noted quite a lot of complaints across a number of providers for this same issue.
Or I get it disconnecting randomly very late at night/early morning but no discernible pattern.
The old Router, which was a BT Hub 4 used to automatically re-set so I wouldn't know of the outage until I looked at the dashboard data. That Router started having other problems that made it unusable hence the replacement with a TP-Link unit.
The TP-Link unit would not automatically re-set and I had to start doing a manual re-boot.
I Contacted TP-Link support who advised some setting changes to IP addresses. That seems to have worked and for weeks I have not had to manually re-boot the Router, even though the dashboard showed I was still getting disconnects late at night/early morning.
Recently though the same problem has started happening, broadband going offline and a manual Router re-boot required however, where previously the disconnect was shown on my dashboard data as being "provider" it is now being shown as "internal" and the overall percentage/number of internal disconnects has increased significantly.
The Mesh wireless remains operative but obviously no internet connectivity.
Does anyone have any ideas why I might be getting more internal disconnects being shown and, if it is correctly being identified as internal, how can I determine what the problem is with my Network?
Any help or ideas would be appreciated.
Would you mind sharing the agent ID please so that I can take a look at your reports to get a sense of what might be going on.
Also, it is always possible that the algorithm is not able to determine what is 'internal' or 'provider' depending on how your network is configured.
We do have a hidden function that allows us to correct this behavior if that is the case. Most times it isn't but it does come up now and then.
Once I've spent some time going over your reports, I can better respond here.
Hi, My agent id is - 126658
I suppose the main thing to note is that this is a change to the behaviour.
I had ended up with two agents one inactive and one active. I was able to reset my original agent (this one) and so deleted the other as they are both looking at the same set-up. I updated the software etc.
Any pointers would be appreciated
I suppose the main thing to note is that this is a change to the behavior.
This is an important observation because OutagesIO is just a tool, like any other. It automated the process of trying to collect information, data, facts that the person can use to try and understand what is going on.
The first thing I notice is no hop2 at this location or it doesn't respond to ICMP. Your hops go from hop1 (presumably your router) to hop3 which is the provider. Because of this, it's hard to know how things are connected.
The average times of your connection problems are not one specific one.
The connection seems to experience problems between 9-10am, again around 12-12:30pm and again around 9:30-10pm.
Note that I'm not 100% sure if your report is showing me the times in your time or mine since the app shows the date/time based on your own timezone. You can confirm this for yourself by looking at those times. Also note that I can see the entire history which is why my result may vary since Community reports show the lat 50 events.
I see pings that average around the 200ms markbut you get spikes that have gone as high as 1721ms.
Your speed tests are also all over the place, even when one hour apart, they are never consistent, always fairly random.
In your speed test settings, you entered 1Mbps as your connection speed. Is this correct? I ask because I see speed test results as high as 7.73Mbps so to me, you have either a purchased speed of a little less than that or a little higher that you're not getting.
In terms of outages, since starting to monitor with this agent, you've had 110 confirmed IP outages.
Looking at your stats, you have around 52% LAN sided problems, 28% provider problems and the last 20% is beyond your provider.
There seems to be something on your local network that is losing connectivity, quite a lot but then so is the provider.
Having said all this, please keep in mind that you may also be experiencing even more down time than what the reports are showing because the agent cannot know if something like a cable gets disconnected or that there are low signal levels depending on how the provider is bringing the service to you. It can only confirm IP outages and is mainly used to know how the provider is doing.
If possible, can you show me a basic diagram of your setup?
Something along those lines?
Hi, Thanks for taking a look at the stats to try and understand what is going on. I agree whole heartedly with your comment that outages is only a tool to provide information to help us determine what is going on. May I add "A really useful tool!"
With regards to the speed testing my bandwith setting was set to 7Mbs but it seems to change to the basic of 1Mbs periodically. I do check and re-set it periodically.
And yes the speed is all over the place going less than 1Mbs on numerous occasions.
I will try and add a diagram of my set-up some time today but lots of stuff going on at the moment so no guarantee.
Basically my set-up is not very complicated.
Connection into the house
Note - Rooms 1, 2 and 3 are downstairs and room 4 is upstairs.
Copper wire connection into a BT 5C Mk 4 Master Socket in room 1 - (We are approx 1.5 miles from the exchange so speed is limited by the distance and getting circa 7Mbs is REALLY good!) No fibre available. Have been offered Fibre to the cabinet (existing copper to the house) with a guaranteed speed of ….......1Mbs!!!
Master socket is already filtered internally and was fitted by Open Reach engineers (Provider engineering firm).
Cable is then split from the Master Socket into separate telephone line and Broadband line - This was also done by the Open Reach engineer.
Telephone line goes to one socket in room 4. We removed the filter after the 5C socket was installed, as advised by the engineers. We have since tried the phone connection with and without the additional filter (makes no difference). We then have wireless telephone set-up through the house so no additional phones or sockets connected.
Cable from Master socket goes to a Broadband master socket in room 2 (also installed by open Reach engineer)
Broadband Master Socket connected via ADSL Cable straight into Combined Modem Router (TP-Link VR900EU Ver 1.0) (in room 2)
LAN Cable from TP-Link router to PC 1, (In room 2)
Lan Cable from TP-Link router to PC2 (In room 2) - Outages tool is installed on this PC and is always on.
Lan Cable from TP-Link router to Mesh Router1-Master (MERCUSYS) (In room 3)
LAN Cable from Mesh Router 1 to Sky Digital TV box (In room 3)
TP-Link settings changed - IP address changed in order to get it to automatically re-connect following loss of broadband connection. Not always re-connecting but more often than not.
Noted more "Internal" disconnects after these settings were changed.
Can you tell me how I can add a picture - Diagram?
I forgot to mention that I use Zone Alarm Extreme security Firewall etc.
Thanks for all the extra information.
When you write a reply, you should see Options and Attachments just below the edit window. Is that showing up?
If so, click on Attachments, Add files and once uploaded, click on the inline option.
I guess the first thing is that it sounds like your provider is getting the Internet to you via DSL. DSL unfortunately, can sometimes be plagued with problems unless the provider has done some extensive upgrades. Telco's usually don't care to invest in their copper infrastructure, mainly interested in trying to earn money using the old technology they are quickly abandoning, meaning, squeezing the last dollar they can get from it.
That said, you're still paying for a service and they need to provider it to some level of reliability, especially if the exchange is only 1.5 miles from your location. Moden DSL should have no trouble with that distance at all.
There are a bunch of possibilities including that they have oversold what they can provide in your area. I think as a next step, if it were possible, I would find another neighbor that is using the same provider and have them install the service as well. This way, you can at least compare your findings to see if there are some correlations.
Sure, if you could get a diagram up, that could help as a next step.
Thanks for the updated comments. Home Network diagram is now attached.
In the UK our providers are pretty much hampered by existing infrastructure and "ownership" BT (British Telecom) used to be pretty much the sole provider so the whole network was owned and managed by them. Slowly, and with the technology changes, new rules have come into place and the network is shared as well as new infrastructure being put in place. Mainly in the more populated areas (large towns/cities). If like me you are rural then regardless of provider you are stuck with using the BT infrastructure and the network repair and management organisation Open reach.
Whenever speed is questioned it always comes down to the distance you are from the exchange and at 1.5miles we apparently are doing exceptionally well to get anything like 7Mbs! Up until a couple of years ago when we had a really good engineer visit we could only periodically get to 1Mbs. He did some tweaks and we suddenly jumped top 7Mbs!! Fantastic.
Neighbours have moved off the BT system to a provision via Dish, more expensive to set up and as we are looking to move house we have not followed suit but will encourage the next owner to do so as the BT timeline for a better service to our location looks to be being managed by their unicorn department!
Thank you for the diagram.
I am surprised that they told you to remove the filter but since I've not personally dealt with DSL in a long time, maybe things have changed.
The most obvious reasons for speed issues are that providers tend to oversell their resources and play a game of what can they get away with until enough customers complain to the point that they are losing money, then finally upgrade and add more bandwidth.
However, speed issues are not always solely because of lack of bandwidth and a lot of other things can affect what appear to be slow speeds or lack of bandwidth. Providers sell you a speed from your location to their nearest point of presence. When you use commercial speed testing tools, they will almost always show you decent speeds because those tools are highly optimized and on the edge of provider networks.
If the problems aren't actually bandwidth, the only way to figure that out since your provider is not going to let you go into their network is to monitor everything possible then try to put the pieces together.
The problems could be anything from a badly configured interface/device by the provider to failing interfaces, switches, all kinds of things which aren't bandwidth related but absolutely affect connectivity.
In this case especially, I would try to find one or two neighbors that would be willing to install OutagesIO then we could look at the compared results. If some obvious things come to light, you and the neighbors could make a case on your providers social page which cannot be dismissed or denied.
All of the above however is based on whether or not the provider is giving you all the facts. If they are telling you 'this is the best you're going to get' then they are being up front. If they are telling you everything looks fine but you know it's not, it is up to you to prove it and hopefully, push them into upgrading in your area with enough complaints.
NOTICE: I think a moderator accidentally edited your last comment instead of replying. To fix this, I am copying their reply into this new comment.
It's good to hear that things have improved some. One issue we see regularly is that providers will make repairs, everything seems to work nicely for a while then slowly degrades again. This seems to be a common thread.
However, I'm now 5 days in and have started to get Outage speed tests showing (Black bars on the graph) yet there is no outage
shown on the outages graph or the "why my internet was down" list in light view.
Can anyone tell me why this might be?
This is a bug in the speed testing algorithm. There is a condition where the agent thinks there is a problem such as latency or a slow down which triggers these speed tests. The black bar is only supposed to show up after an outage.
The idea behind running the speed test after an outage is that over time, it might show a trend. It's not all that useful since speeds are constantly changing and typically, right after an outage, speeds are way back up at least for a while.
The reports are all about trends, it's what you see over longer periods of time but when you don't see an outage yet see the black bars, that's a bug we've been trying to iron out.