Hosts availability discrepancies

[Babar_Qayyum]

Dear All,

I wanted to clear my concept of infrastructure monitoring. Today, I noticed that one of the servers went offline at 10:30 AM and came back at 10:54 AM.

In the hosts' dashboard, one was grayed-out and other was active.

Same Host

The downtime/offline time was 24 minutes but we can see that the offline time of the unavailable host keeps increasing even though the server is available.

Offline State

Also, there is no connectivity on the graph between offline and uptime with the unavailable time.

Online State

Regards,

Babar

[skrystosik · Accepted solution]

If you have new host not old one back active this means that agent was reinstalled all even host recreated. We have the same issue with AWS deployments where hosts are destroyed and started again. Right now there is no option for linking them for continuous data set.

Data are missing because during blank there was no monitoring at all.

Sebastian

[Babar_Qayyum]

Hello @sebastian k.

I am not sure whether the system was restored or simply restarted. If it was simply restarted then I am wondering about the hosts alerting.

Regards,

Babar

[skrystosik]

If this was only restart you should see the same host online. There is special ID generated on host after installing agent. This is why after each restart data are sent to the same host tab in dynatrace.

Sebastian

[Babar_Qayyum]

Hello @sebastian k.

Where can I see this ID to distinguish between both host agents (offline/online)?

Regards,

Babar

[skrystosik]

Host id is always in URL when you go to host page.

Sebastian

[Babar_Qayyum]

Hello @sebastian k.

I could not find the Host ID in both hosts (Online and Offline).

Online host metadata

Offline host metadata

Regards,

Babar

[skrystosik]

URL 🙂 not properties. There is ID of entity that you are looking at.

Sebastian

[Babar_Qayyum]

Hello @sebastian k.

🙂 thank you for pointing me to the right location.

I can see that both hosts (offline and online) have different IDs. So can I assume that this was a restoration of the VM instead of a restart?

• #newhosts/hostdetails;id=HOST-A9A81E6F3C836130;
• #newhosts/hostdetails;id=HOST-39D1C4FC40ACC8D8;

Regards,

Babar

[skrystosik]

You're correct 🙂 In general if you will see the same host as new instance it always means that new ID was created. This happen mostly during destroy and recreate of host.

Sebastian

[Babar_Qayyum]

Hello @sebastian k.

Thank you for sharing this information. How can we avoid false-positive availability alerts in this situation?

Regards,

Babar

[skrystosik]

I was posting RFE for ability to manual define such host ID during installation of agent for such purposes but for now, no response.

https://answers.dynatrace.com/spaces/483/dynatrace-product-ideas/idea/217428/rfe-support-for-telling...

Sebastian

[Babar_Qayyum]

Hello @sebastian k.

I voted RFE, and I hope Dynatrce should bring a solution to this issue as we are planning to replace a tool which is only doing a host monitoring.

Regards,

Babar

[Christopher_Cha · Accepted solution]

I just want to add on to what Sebastian has mentioned. Usually if it's a static host, then upon uninstallation of the agent, there would still be some static remnants left behind; these are used when we reinstall the agent, and the agent will recognize that this was previously monitored so data continuity will be maintained. It's a totally different story if it's cloud deployment where hosts can be destroyed and "respawn" as completely new host.

[Babar_Qayyum]

Hello @Christopher C.

Thank you for your comments. We have a managed setup with a lot of virtualized setup. If system administrators restore the VM at any stage then there will be two different host IDs as we discovered in the current issue.

Regards,

Babar