Track Internet disconnections, provider outages with historical data, and automated speed testing.
For Windows, Linux, ARM64, ARMa7. Learn more by visiting www.outagesio.com
Notice: If you created an account on app.outagesio.com, simply use the same credentials to log in here.
No Data: Outages, Avg Time, Network Stats, Speed Test Results
-
I split this topic up from the other so that we could better keep track of different posts.
I quickly looked at the agent and don't see any problems but I do see that it is disconnected at the moment.
One problem we are seeing is members restarting their PC or the agent service every time they see it disconnected.
If you can reply here once it's back up, I'll monitor it in real time and see if I notice anything unusual.
Good idea to split topic. Different user - who by the way just had his first week at work in a new job. Wehee!!Yes, this laptop has led - and will continue to lead - an on/off lifestyle. I'm no fan of powerhungry devices running 24/7/365. Thus the wish for a Hardware Agent, once I see proof of the system's ability to detect and record an outage.
But today - february 6, 9:00 CET (that's central Europe) to 22:00, I'll leave the machine running and hope you read this in time to take a check.
NB: System info update. ISP decided to double bandwidth at no cost. Now it's 200/200 MBit :)
//Martin
-
Hmmm - poking around on the web-interface I came across what may be helpful to you.
Hovering the mouse OVER the "Outages avg time" graph, something flickered as I moved the mouse pointer around. I managed to grab it:
https://i64.servimg.com/u/f64/19/68/15/14/outage10.png" />Removing the pointer from the area left it completely blank - nothing at all to be seen.
Following this, I logged out and back in again using the demo account, to see how the interface looked on that one:
https://i64.servimg.com/u/f64/19/68/15/14/outage11.png" />Same browser, same browser window, same PC - but different results from different accounts.
Would this be a clue?
I'm running Firefox v85.0.1 on Win10/64
//Martin
-
Hi there,
Sorry it took so long to reply. We've been swamped with work lately.
In the 'Outages average' that you have posted an image of, the reason nothing is showing is because there has been no data to report. Meaning, this connection does not appear to have suffered any IP outages though there are plenty of Inactive/Disconnected periods.
It appears that you are comparing our demo reports and your own which would not be the same at all.
Heartbeat, Speedtest & Ping: Showing data/OK
Outages, AvgTime & NetworkStats: Empty
This usually means that there is something else, not IP specifically but something else that is causing the disconnections like a cable going bad, or an interface or router, bad config, any number of things. The agent will always report IP based outages but it cannot know what caused disconnections.
Trying to get confirmation my setup is OK, I've then tried poking the system:
From what we can see on our end at least, things look fine.
Just one small point to make. While the service can identify most IP outages, in many cases, there is an element of human intervention that is required. Meaning that like any tools, you have to use it as just that, something that gives you leads and information that can help you to dig further.
Shut down fiber connector box (that should kill the WAN). Checking via cellphone I see the agent becoming 'Inactive'.
Restore power to fiber connector box. Validate Agent comes back 'Active' via cellphone-link.
In your dashboard, if you see Inactive, and eventually (after 30 minutes or so), Disconnected, it means the agent was no longer able to communicate with the OutagesIO network. However, when it does come back, assuming the Windows service or hardware agent were not stopped, if there was an outage, it will send its report as soon as it can communicate with our network again. If the agent comes back online after an Inactive/Disconnected but it doesn't send an outage report, it means there was no IP outage but something else that happened.
Shut down router (that should kill the LAN). Checking via cellphone I see the agent becoming 'Inactive'.
Restore power to router. Validate Agent comes back 'Active' via cellphone-link.
This I would need to better understand your network topology. Any disconnection of the network would definitely cause an Inactive and if longer than 30 minutes or so, a Disconnected status in your OutagesIO dashboard. However, depending on your network topology, the agent may or may not consider the local outage or loss of the gateway as an outage. It is mainly trying to determine if your provider is experiencing problems.
Shut down Agent (That should register as inoperative Agent or something similar). Checking via cellphone I see the agent becoming 'Inactive'.
restart Agent. Validate Agent comes back 'Active' via cellphone-link.
This is correct, shutting down the agent would only cause an Inactive/Disconnected to show in your OutagesIO dashboard.
The agent is only looking for IP outages but the overall solution can also tell you when your location is simply not able to communicate over the Internet. One also has to take into account that different traffic can take different paths source to destination.
For example, the agent may not be able to reach the OutagesIO network but you are still able to reach Facebook on your browser. This implies that something along the path (hops) taken to the OutagesIO went down but the path to Facebook wasn't affected.
Having said that, the number one purpose behind OutagesIO is to monitor your provider but all network providers worth their salt have multiple incoming and outgoing paths. Therefore, one thing could still continue to communicate while another might not, it depends on the network path that those packets took.
Another way to really help determine these things is to have one or two neighbors use the service as well. You'll notice your dashboard has a little map. This is to show you other agents in the area that are experiencing similar problems. While the map will only show you in near real time those events, the other information you all have can easily be correlated by comparing your reports.
I hope this helps to explain a little better how OutagesIO works and what you should be looking for.
We are happy to try and help if you want to keep this open and post new findings.
-
I'm not done with this.
That said, i have little time (got a new job last week) to fiddle around with PCs and software agents, so I've bit the bullet and ordered a hardware Agent. It's no fortune afterall, if you disregard VAT, toll and various import fees.
Given time and opportunity, I'd like to execute - and meticulously document - a test of the complete system's ability to detect and record outages to the extent of what I'm capable to cause.
* Router off: No WAN connection, no DHCP, nothing. Note here that LAN switch != router.-
Router with no WAN connection: DHCP, local nodes like printer and NAS available.
-
I can't simulate an ISP error. We'll have to wait for that to happen by itself.
Note here that my setup is with a static IP and all internal LAN addresses administered by my own local router. ISP delivers a hole, and I pay to always have the same hole address. I run the firewall on that router, with the kind and effective assistance of the nice folks of https://www.nic.cz/ (specifically https://www.turris.com/en/omnia/overview/).
Anyways, hproper Hardware agent is on its way. It'd better prove its worth.
-
-
Someone posted recently that they were seeing substantial differences in speed test results between a hardware and software agent at the same location. This is something we are investigating but have always said that speed testing is kind of a dark art :).
I'm not sure what 'proving its worth is' but as usual, you must take the results with a grain of salt. If your main objective is to critique the service, I'm sure you'll find plenty to comment about since we are a relatively new venture doing our best to keep up with everything needed to offer the best service we possibly can. Some people prefer to simply say 'your service doesn't work' while the ones that give us good input and feedback actually help us further develop the service.
Trying to fully automate something that usually always takes human intervention is quite the challenge but with member feedback, we can always improve every aspect of the service.
While there are fairly 'standard' networks, there are also many cases where one network doesn't quite work in the same way as another specifically because of the switches, firewalls, routers being used and how they are connected.
This is the important part. Depending on your setup, the agent may or may not be able to recognize IP outages that are on the LAN but it should always be able to determine IP outages on the next network.
All of that said, the best way to connect OutagesIO would be as close to the provider as possible because that is the main point of the service, to monitor your provider, not your LAN. Keep in mind that if there are several private IP hops from your network to your providers, the algorithm may not be able to figure out where your LAN ends and where your provider starts. However, we do have an override that we can use if you find that this is the case.
Sometimes, it helps when members post a simple diagram of their network so that it is a little easier for us to help.
-
HW agent arrived. ID 127572.
I'm now waiting for it to document an outage - which is its purpose, afterall…
-
Great.
Now, just keep in mind that inactive status do not equate to outages.
There are posts about statuses and notifications in the forums.
The agent will always log outages with the provider once you have any.
Beyond are informational only.
-
Someone posted recently that they were seeing substantial differences in speed test results between a hardware and software agent at the same location. This is something we are investigating but have always said that speed testing is kind of a dark art :).
All of that said, the best way to connect OutagesIO would be as close to the provider as possible because that is the main point of the service, to monitor your provider, not your LAN.
With HW agent now installed, I've opted for LAN connection.
And yes; describing my LAN setup obviously let you assess its impact on detection efficiency. WillDo:
* Draw graphic of LAN- Post graphic on forum.
Now, wait…
-
keep in mind that inactive status do not equate to outages.
Given my power supply to the household rarely (never) faults, and the HW agent's operation depends on only that and the router's state (I get emails prior to reboots - the nice folks at [](https://www.turris.com/en/omnia/overview/)[https://www.turris.com/en/omnia/overview/](https://www.turris.com/en/omnia/overview/) made that possible), I'd expect to see nothing but ISP outages detected. -
I'm sure I don't have to mention it but it might help someone else that finds this post.
If/when there seems to be something happening with the Internet, please make sure the agent continues to run non stop.
A lot of people see the agent inactive or disaconnected in their dashboard and restart the agent which causes it to lose any actual outage data it might have been tracking.
The agent will always track outages but sometimes it takes some human intervention to decypher the results later. For example, you might still have Internet access but the agent is unable to reach the OutagesIO network. Things of that nature become clearer as data builds up and you start seeing problem points etc.
We'll be happy to help out once you start seeing this kind of information coming in. Having a neightbor or two that are using the same provider monitoring also is an excellent way to confirm provider issues.
