Information and Support

As an update, there is another thread in the forums where someone has around 30 agents and some random problems were coming up.

After months of testing, we were finally able to isolate the problem and we thought it was fixed.

Additional logs provided to us using the latest version shows that the problem is still there in some form or another.

The agent uses a multithreaded method and it seems that in some cases, one of the threads can spiral out of control causing the agent to miss sending some data now and then.

What's probably happened from looking at your reports is that the agent missed sending pings because the rest of the data looks consistent while the missing pings have no associated missing communications.

In other words, when those pings were missed, the gaps, the agent was still communicating which tells us it failed to send the pings in those times.

When you see gaps, you can always look at your outages to see if there was an outage at the same time. If so, you can discount the missing ping/heartbeat as it makes sense since the agent was not able to communicate. Otherwise, you can attribute the gaps to the problem we are now working on and will release an update once tested.

The biggest challenge with offering online services is that when folks come across problems, they usually prefer to blame the service and tell others how bad it is.

The fact is, if people took the time to let the service know about the problems, they will always do everything possible to fix any problems to keep hard-earned members.

Thank you for pointing this out.

No problem.

There is a bit of a learning curve but not too much. We've tried to make it as automatic as possible but like all tools, it does take a little reading and learning to understand how to use it.

The www.outagesio.com site is mainly just our marketing site, it's where search engines and folks searching for a service like ours can learn about us.

This site is for support but also contains quite a lot of information from us and others who have posted similar questions that can help others.

The app.outagesio.com site is the service itself. It's a 100% custom built service which is why all three of these things aren't all in one service. It simply wasn't possible.

My suggestion is that if you have an agent installed at this point, just let it run for a while. Data will start showing up as problems occur.

Once it does, review the pages help and if you need more clarification, just post here including some details (not personal) and we'll try to help.

As explained earlier, when you see disconnections but no outages, it means anything but IP outages. Those links I shared explain what some of the causes could be, bad or slowly failing cable, firmware needing an update in your firewall or router or modem, or your provider has a signal level problem at the street level and they aren't acknowledging the problem.
The only way to know more in your case would be to get one or two neighbors to install an agent also so that we can correlate their data.

To get more info, you could always upgrade your reports for 30 days and that will show you historical data and other information that may or may not help. It may not help since you seem to be experiencing problems which aren't specifically IP related so the agent cannot log those, it can only tell you that's what seems to be happening.

You can know this because of what you aren't seeing, IP outages that would be logged by the agent.

@SBK Thanks Ed. Funny we mentioned the lower prevalence of runtime bugs in code written in the functional programming paradigm, then a runtime bug happens, what are the odds?

Aside from this little hiccup, the service has been great though. Appreciate the quick fix.

Most times, you'll know if even your own local gateway goes down but a lot of that has to do with how a network is set up.

Some folks have multiple internal devices like a wireless router connected to their own router which is then connected to their providers.

There is no way for the agent to know these things so to us, it's expected to work on a 'traditional' network which is usually the providers router/modem only.

That said, for companies that do have multiple internal gateways, we have a feature we can manually enable that tells the reports side where the local network ends and where the provider starts otherwise, those folks get reports of the LAN being down when it's actually the provider or vise versa.

Hard to have a tool that does it all automatically but we keep working on it :).

@kmichailg

Yes, this confirms the change we decided to apply: we want to avoid cookies usage in the login phase at least.

That error is the effect of that change.

Please inform of any other situation where the login for some reason is affecting the normal behavior of the app