Do you need to know how your GIS environment is performing, and how it impacts your end-users’ experiences?
Geocortex Analytics has the tools you need to efficiently monitor your activities, avoid unnecessary interruptions, and keep your users happy.
In the first of our Geocortex Analytics themed Geocortex Tech Tip videos, which we’ll be running over the next month, we’ll show you to manage alarms when a resource in your environment is unavailable or experiencing performance issues.
This tech tip will help you to set alarm thresholds and notification settings to best suit your organization’s GIS needs, as well as how to interpret triggered alarms.
“Hi, my name is Stephen and I am on the Analytics team. Today, we are going to take a look at alarms in Analytics including how to configure alarms, how to interpret alarms, and how to receive alarm notifications. So, let’s jump in!
Alright, let’s get started! Before we get to dive into alarm configuration, let’s first see what alarms look like in Analytics Reports.
I’m currently on the ‘Status’ > ‘System’ page, and here we have the ‘Active Alarms’ panel. Now, this panel provides information on all the resources that Analytics is monitoring that are currently either unavailable or having performance problems. From an administrative standpoint, this is really important because it tells you which resources require your attention right now.
So, here we distinguish between two different types of states for resources. The first one is this red critical error that indicates that Analytics believes this resource is down or unavailable, or maybe just simply cannot be reached by Analytics, and so this something that you want to investigate right away.
Alternatively, we also have this orange warning label that indicates that the resources aren’t performing optimally with some kind of performance issue. For instance, we have an ‘LA_Layers’ map service here, and it was a response time alarm that was triggered when it took 11 seconds. So, this probably longer than we’d ideally want to see.
Going back to this first ‘Request Failure Alarm,’ I’m going to show you what it looks like when you expand the details there. We had a request failure alarm, which means that we tried to make a request to this image service and it either failed or — in this case — it was a passing error on the response.
This happens a lot when the ArcGIS server does actually send back a response, but that response contains an error message.
We can see from the response here, that the error message was ‘Could not access any server machines. Please contact your system administrator.’ For an image service, it might mean that is a disconnect between the Geo database and the ArcGIS Server or maybe something similar, so that will need to be investigated.
Then also in this detailed panel right here, we show you when it was last checked. So, Analytics is going to continue checking a resource, until it discovers that it’s healthy and then that resource will be removed from this panel, but we can see that it’s been checked relatively recently, and it was still down, so it’s likely still a problem. We also show when it entered the current state, so this is when it entered a broken state and that was about a day ago, so we know approximately know for how long this resource has been broken.
Lastly, we show you within the last 24 hours how often it was down. So, if a resource keeps going up and down, it’s going to look like a spiky graph, but if it was just permanently offline — as the service was — then you’re just going to see a very straight line.
Let’s shift gears and talk about alarm configuration. Another resource that is showing up here is this host server that has a maximum memory usage alarm, and the RAM utilization is about 83% on that machine, which is above the threshold of the default threshold of 80%. But maybe we don’t care, unless a host server goes above 95% RAM utilization, and in that case, we would need to go and configure that alarm.
So, we’re going to go to the ‘Server,’ from there were going to go to ‘Configuration’ and from here you’ll see that there’s an ‘Alarms,’ tab and this contains all the alarms that can be configured for that resource. In this case, we talked about the ‘Memory Usage Alarm’, and you can see the current settings. So, it’s set to 80% as if it goes above 80% that’s when it’s going to trigger and the number of occurrences. If that was set to ‘2,’ it would need to be above 80% on two consecutive collection operations in order for the alarm to trigger.
Now, these options are greyed out, because the ‘Use Default Settings’ toggle is currently enabled and what this means is that there a default settings for each resource type for each alarm and so there a default server settings for the memory usage alarm that is set to 80 and those are applied to all the servers by default.
I have two options right now, I can either change the default settings to 95% which would then impact all the servers, or I could override the default settings on this one server if I only wanted that to have a threshold of 95%.
In this case, I actually want to set that 95% threshold, and have it applied to all the servers. I’ll cancel out of here. I’m going to go to ‘Servers,’ and this page also has an ‘Alarms’ tab that has the default settings for all of those different alarm types. I’m going to go to ‘Memory Usage Alarm’ again, and then I’m going to set it to 95% and it saved.
Now if I go back here, you’ll notice that it was set to 95% because it is inheriting those default settings.
Let’s say that for this machine I actually care more about how much RAM utilization there is. I’m going to disable overwrite default settings and say 50%. I want to be notified as soon as memory usage is over 50% on this one particular machine and hit save.
Now let’s go back to the system page and the active alarms panel to see what that looks like.
As you can see the alarm for the other server that we were looking at before, ‘QA-AGS-1071’ one is gone and instead we now have an alarm that’s triggered for this other, ‘SBATES-VM2019’ machine at 63% because we lowered the threshold to 50%, which is exactly what we were looking for.
So, the ‘Active Alarms’ panel is a good place to check on the state of your GIS environment, but you really want to be notified immediately when a resource goes down or starts to experience performance issues. And for this, you want to set up email and or SMS text notifications.
To do this, we need to go back to ‘Configuration’ under the ‘System’ page, there’s an ‘Alarm Recipients’ tab, this recipient tab should contain all the information of the individuals that wish to receive either an email or text message when there’s something wrong with the system.
Right now, I’m going to add a new recipient, and I’m going to add myself and I’d also add a phone number if I wanted to be contacted that way. At the bottom here, there is an option ‘Receive Notifications for All Resources’. If I were to check this for email, for instance, this user would receive all notifications for every single resource in the system, no further configuration will be required.
But, in this case, I kind of want to fine-tune it a little bit and only add myself as a recipient to very select alarms, so that I’m not inundated with emails. I’m going to hit save.
Also worth noting is that in order for Analytics to send out emails, you need to go into ‘Email Settings’ to actually configure SMTP Server settings and similarly you need to go into ‘SMS Settings’ to provide you a Twilio account details if you wish to send out text messages.
Now that I’ve added myself as a recipient, I’m going to go back to the original server that we were looking at. Go to the ‘Alarms’ tab, and I want to be notified as soon as that server is unavailable as denoted by this alarm.
I’m going to edit that and under ‘Email Recipients,’ I’m going to select myself. By hitting save, I’ve now added myself to this configuration and I’ll be sent an email as soon as the alarm is triggered.
There’s a second way that you can add a recipient to an alarm which is basically in-line so that you don’t need to go to the recipients page, and this is, I can just type out a user. They don’t exist, that’s fine I’m going to click ‘Add New Recipient,’ I’m going to fill in the details of that user, hit save. Immediately they’ve been added, and they will also be added to the global recipient list if you want to add them to a different alarm.
That concludes this Tech Tip. Hopefully, you’ll have a better understanding of how alarms work and how to configure them. Thanks for joining me and I’ll see you next time!”
Want to see more of Geocortex Analytics in action? Click the button below to learn more about the product, or to schedule a personal demonstration.